Z80

ASSEmBLY lAnGUAGE
PROGRAmminG

Z80
ASSEmBLY LAnGUAGE
PROGRAmminG

Lance A. Leventhal

Osborne/McGraw-Hili
Berkeley, California

Published by
OSBORNE/McGraw-Hili
630 Bancroft Way
Berkeley, California 94710

U.S.A
For information on translations and book distributors outside of the U. S. A. ,
please contact the publisher at the above address.

56 789

DODO

8 76 54 3 2

ISBN 0-931 988-21- 7

Copyright

1979 by McGraw-HilL Inc.

All rights reserved. Printed in the United States of America. No part of this
publication may be reproduced. stored in any retrieval system, or transmitted
in any form or by any means, electronic, mechanical, photocopying, recording
or otherwise, without the prior written permission of the publishers.
Cover design by K. L. T. van Genderen.

This book is dedicated to my colleagues at the Society for Computer Simulation - Romeo Favreau. Natalie Fowler. Alexander McKenna. John McLeod.
Stanley Rogers. and Chip Stockton.

ACKNOWLEDGMENTS
The author would like to acknowledge the following people:
Mr. Curt Ingraham, Ms. Mary Borchers, and Ms. Janice Enger of Osborne/
McGraw-Hili, who made many corrections and suggestions; Mr. Winthrop
Saville of Sorrento Valley Associates, who provided assistance and examples; Mr. Tom Littlefield of Littlefield/Smith Associates, who provided
reference material; Ms. Marielle Carter of Sorrento Valley Associates, who
typed some of the material; Mr. Stanley Rogers of the Society for Computer
Simulation, who has continued to suggest improvements in the author's writing style; and his wife Donna. for her patience and understanding throughout
the writing of this book.
Others who provided assistance and suggestions were Mr. Colin Walsh, Mr.
Gary Hankins, Mr. Romeo Favreau, Mr. David Bulman, Ms. Kati Bulman. Mr.
Robert Turner, Mr. Irv Stafford, Mr. John Burgar, Mr. Ferenc Montvai-Lako,
and Mr. Warren McKenna. Other students and colleagues also helped to keep
the author on the right track.
The author, of course, bears responsibility for any remaining errors, misconceptions, and misinterpretations.

Contents
Chapter

Page
Introduction to Assembly Language Programming
How This Book Has Been Printed
The Meaning of Instructions
A Computer Program
The Programming Problem
USing Octal or Hexadecimal
Instruction Code Mnemonics
The Assembler Program
Additional Features of Assemblers
Disadvantages of Assembly Language
High-level Languages
Advantages of High-level Languages
Disadvantages of High-level Languages
High-level Languages for Microprocessors
Which Level Should You Use 7
How About the Future?
Why This Book?
References

1-1
1-1
1-1
1-1
1-2
1-3
1-4
1-5
1-6
1-6
1-7
1-7
1-8
1-9
1-10
1-11
1-11
1-12

Assemblers
Features of Assemblers
Assembler Instructions
Labels
Assembler Operation Codes (Mnemonics)
Pseudo-operations
The Data Pseudo-operation
The Equate (or Define) Pseudo-operation
The Origin Pseudo-operation
The Reserve Pseudo-operation
linking Pseudo-operations
Housekeeping Pseudo-operations
Labels with Pseudo-operations
Addresses and the Operand Field
Conditional Assembly
Macros
Comments
Types of Assemblers
Errors
Loaders
References

2-1

vii

2-1
2-1
2-2

2-4
2-4
2-5
2-6

2-7
2-7
2-8
2-8
2-9
2-9
2-11
2-11
2-13

2-14
2-14
2-15
2-15

Contents (Continued)
Page

Chapter

The Z80 Assembly Language Instruction Set

CPU Registers and Status Flags
Z80 Memorv Addressing Modes
Implied
Implied Block Transfer with Auto-Jncrement/Decrement
Implied Stack
Indexed
Direct
Program Relative
Base Page
Register Direct
Immediate
Abbreviations
Instruction Mnemonics
Instruction Object Codes
Instruction Execution Times
Status
Instruction Descriptions
8080A/Z80 Compatibility
Zilog Z80 Assembler Conventions
Assembler Field Structure
Labels
Reserved Names
Pseudo-operations
Examples
Labels with Pseudo-operations
Addresses
Conditional Assembly
Macros

3-1
3-2
3-4
3-5
3-7
3-8
3-10
3-11
3-12
3-13
3-14
3-15
3-18
3-21
3-21
3-21
3-21
3-43
3-164
3-170
3-170
3-170
3-170
3-170
3-171
3-172
3-172
3-174
3-174

Contents (Continued)
Chapter

Page
Simple Programs
General Format of Examples
GUidelines for Problems
Program Examples
Ones Complement
8-Bit Add iliOn
Shift Left One Bit
Mask Off Most Significant Four Bits
Clear a Memory Location
Word Disassembly
Find Larger of Two Numbers
16-Bit .Addition
Table of Squares
16-Bit Ones Complement
Problems
Twos Complement
8-Bit Su btractlon
Sh ift Left Two Bits
Mask Off Least Significant Four Bits
Set a Memory Location to All Ones
Word Assembly
Find Smaller of Two Numbers
24-Bit .Addition
Sum of Squares
16-Bit Twos Complement

4-1
4-1
4-2

Simple Program Loops

Examples
Sum of Data
16-Bit Sum of Data
Number of Negative Elements
Find Maximum
Justify a Binary Fraction
Problems
Checksum of Data
Sum of 16-Bit Data
Number of Zero. Positive. and
Negative Numbers
Find Minimum
Count 1 Bits

5-1

IX

4-3
4-3
4-4
4-6
4-6
4-7
4-7
4-9
4-11
4-12
4-14
4-15
4-15
4-15
4-16
4-16
4-16
4-16
4-16
4-16
4-17
4-18

5-3
5-3
5-6
5-9
5-11
5-14
5-17
5-17
5-17
5-18
5-18
5-18

Contents (Continued)
Page

Chapter

Character-coded Data
Examples
Length of a String of Characters
Find First Non-blank Character
Replace Leading Zeros with Blanks
Add Even Parity to ASCII Characters
Pattern Match
Problems
Length of a TeletypeWriter Message
Find Last Non-blank Character
Truncate Decimal String to Integer Form
Check Even Parity In ASCII Characters
String Comparison

6-1
6-2
6-2
6-8
6-11
6-13
6-16
6-19
6-19
6-19
6-20
6-20

Code Conversion
Examples
Hex to ASCII
Decimal to Seven-Segment
ASCII to Decimal
BCD to Binary
Convert Binary Number to ASCII String
Problems
ASCII to Hex
Seven-Segment to Decimal
Decimal to ASCII
Binary to BCD
ASCII String to Binary Number
References

7-1
7-1
7-1
7-3
7-8
7-10
7-11
7-13
7-13
7-13
7-14
7-14
7-14
7-15

Arithmetic Problems
Examples
MUltiple-Precision Addition
Block Move
Decimal Addition
8-Bit Binary Multiplication
8-Bit Binary Division
Self-Checking Numbers Double and
Double MOD 10
Problems
Multiple-PrecIsion Subtraction
Decimal Subtraction
8-Bit by 16-Bit Binary MultiplicatIOn
Signed Binary Division
Self-Checking Numbers Aligned 1,3,7 MOD 10
References

8-1
8-1
8-1
8-4
8-5
8-8
8-12

6-21

8-17
8-25
8-25
8-25
8-26
8-26
8-27
8-28

Contents (Continued)
Chapter

Page

Tables and Lists

Examples
Add Entry to List
Check an Ordered List
Remove Element from Queue
8-Bit Sort
USing an Ordered Jump Table
Problems
Remove an Entry from a List
Add an Entry to an Ordered List
Add an Element to a Queue
16-Bit Sort
USing a Jump Table with a Key
References

9-1
9-1
9-1
9-5
9-8
9-10
9-14
9-16
9-16
9-17
9-17
9-18
9-18
9-19

10

Subroutines
Subroutine Documentation
Examples
Hex to ASCII
Length of a String of Characters
Add Even Panty to ASCII Characters
Pattern Match
Multiple-PrecIsion Addition
Problems
ASCII to Hex
Length of an ASCII Message
Check Even Panty In ASCII Characters
Stnng Companson
Decimal Subtraction
References

10-1
10-2
10-2
10-3
10-6
10-9
10-12
10-16
10-19
10-19
10-19
10-19
10-20
10-20
10-22

11

Input/Output
Timing Intervals (Delays)
Delay Routines
Example
Delay Program USing Accumulators
Simple I/O Devices
The Z80 Parallel Input/Output CirCUit (PIO)
PIO Mode Control
Configunng the PIO
Z80 Input/Output InstructIOns
Examples
A Pushbutton SWitch
A Toggle SWitch
A Multiple-Position (Rotary. Selector. or
Thumbwheel) SWitch
A Single LED
Seven-Segment LED Display

11-1
11-8
11-8
11-9
11-9
11-11
11-11
11-15
11-17
11-18
11-22
11-22
11-28

xi

11-33
11-40
11-43

Contents (Continued)
Page

Chapter
11 (Cont.)

12

Problems
An on-off Pushbutton
Debouncing a Switch In Software
Control for a Rotary Switch
Record Switch Positions on lights
Count on a Seven-Segment Display
More Complex I/O Devices
Examples
An Unencoded Kevboard
An Encoded Keyboard
A Digital-to-Analog Converter
Analog-to-Digital Converter
A Teletypewriter (TIY)
The ZSO Serial Input/Output Device (SIO)
Examples
Teletypewriter I/O via a USART
Standard Interfaces
Problems
Separating Closures from an Unencoded
Keyboard
Read a Sentence from an Encoded Keyboard
A Variable Amplitude Square Wave Generator
Averaging Analog Readings
A 30 Character-per-Second Terminal
References
Interrupts
ZSO Interrupt Svstem
Non-Maskable Interrupt
ZSO Interrupt Modes
ZSO/SOSO Interrupt Compatibility
PIO Interrupts
Examples
SIO Interrupts
Interrupt Examples
A Startup Interrupt
A Keyboard Interrupt
A Printer Interrupt
A Real-Time Clock Interrupt
A TeletypeWriter Interrupt
More General Service Routines
Problems
A Test Interrupt
A Keyboard Interrupt
A Printer Interrupt
A Real-Time Clock Interrupt
A TeletypeWriter Interrupt
References

xii

11-55
11-55
11-55
11-55
11-56
11-56
11-57
11-60
11-60
11-69
11-72
11-76
11-Sl
11-S9
11-9S
11-9S
11-103
11-103
11-103
11-103
11-104
11-104
11-104
11-105
12-1
12-2
12-3
12-4
12-5
12-6
12-S
12-10
12-12
12-12
12-14
12-17
12-20
12-26
12-30
12-31
12-31
12-31
12-31
12-31
12-31

12-32

Contents (Continued)
Chapter

13

Page

Problem Definition and Program Design

The Tasks of Software Development
Definition of the Stages
Problem Definition
Defining the Inputs
Defining the Outputs
Processing Section
Error Handling
Human Factors
Examples
Response to a SWitch
A SWitch-Based Memorv Loader
A Verification Terminal
Review of Problem Definition
Program Design
Flowcharting
Examples
Response to a Switch
The Switch-Based Memory Loader
The Credit-Verification Terminal
Modular Programming
Examples
Response to a Switch
The Switch-Based Memory Loader
The Verification Terminal
Review of Modular Programming
Structured Programming
Examples
Response to a Switch
The Switch-Based Memorv Loader
The Credit-Verification Terminal
Review of Structured Programming
Top-Down Design
Examples
Response to a SWitch
The SWitch-Based Memorv Loader
The Transaction Terminal
Review of Top-Down Design
Review of Problem Definition and Program Design
References

xiii

13-1
13-1
13-3
13-3
13-4
13-4
13-5
13-5
13-6
13-6
13-6
13-8
13-11
13-15
13-16
13-17
13-19
13-19
13-20
13-22
13-26
13-28
13-28
13-28
13-28
13-30
13-30
13-36
13-36
13-36
13-38
13-43
13-44
13-45
13-45
13-46
13-47
13-49
13-49
13-50

Contents (Continued)
Page

Chapter

14

15

16

Debugging and Testing

Simple Debugging Tools
More Advanced Debugging Tools
Debugging with Checklists
Looking for Errors
Debugging Example 1 Decimal to Seven-Segment
Conversion
Debugging Example 2: Sort Into Decreasing
Order
Introduction to Testing
Selecting Test Data
Testing Example 1 Sort Program
Testing Example 2: Self-Checking Numbers
Testing Precautions
Conclusions
References

14-1
14-1
14-8
14-10
14-11

Documentation and Redesign

Self-Documenting Programs
Comments
Commenting Example 1 Multiple-Precision
Addition
Commenting Example 2: Teletypewriter Output
Flowcharts as Documentation
Structured Programming Languages as
Documentation
Memory Maps
Parameter and Definition Lists
Library Routines
Library Examples
Library Example 1 Sum of Data
Librarv Example 2: Declmal-to-Seven-Segment
Conversion
Librarv Example 3: Decimal Sum
Total Documentation
Redesign
Reorganizing to Use Less Memory
Major Reorganizations
References

15-1
15-1
15-2

Sample Projects
Project :#:1: A Digital Stopwatch
Project :#:2: A Digital Thermometer
References

16-1
16-1
16-15
16-29

14-16
14-21
14-27
14-28
14-29
14-29
14-29
14-30
14-31

15-4
15-5
15-7
15-7
15-7
15-8
15-10
15-10
15-10
15-11
15-12
15-13
15-14
15-15
15-16
15-18

xv
xvii

Index of Instruction Descriptions

Index

xiv

Chapter 1
INTRODUCTION TO ASSEMBLY
LANGUAGE PROGRAMMING
This book describes assembly language programming. It assumes that you are
familiar with An Introduction To Microcom~uters: Volume 1 - Basic Conce~ts
(particularly Chapters 6 and 71. This book does not discuss the general features of
computers, microcomputers, addressing methods, or instruction sets; you should
refer to An Introduction To Microcomputers: Volume 1 for that information.

HOW THIS BOOK HAS BEEN PRINTED

Notice that text in this book has been printed in boldface type and lightface type.
This has been done to help you skip those parts of the book that cover subject
matter with which you are familiar. You can be sure that lightface type only expands on information presented in the previous boldface type. Therefore, only read
boldface type until you reach a subject about which you want to know more, at which
POint start reading the lightface type.

THE MEANING Of INSTRUCTIONS

The instruction set of a microprocessor is the set of binary inputs which produce
defined actions during an instruction cycle. An instruction set is to a microprocessor
what a function table is to a logiC device such as a gate, adder. or shift register. Of
course, the actions that the microprocessor performs In response to the Instruction inputs are far more complex than the actions that combinatorial logic devices perform In
response to their Inputs.
An instruction is simply a binary bit pattern - it must be
BINARY
available at the data inputs to the microprocessor at the
INSTRUCTIONS
proper time in order to be interpreted as an instruction. For example. when the Z80 microprocessor receives the 8-bit binary pattern 10000000 as the
input dUring an instruction fetch operation, the pattern means:
"Add the contents of Register B to the contents of the Accumulator"
Similarly, the pattern 00111110 means:
"Load the Accumulator With the contents of the next word of program memory"
The microprocessor (like any other computer) recognizes only binary patterns as instructions or data; it does not recognize words or octal. deCimal. or hexadeCimal numbers.

A COMPUTER PROGRAM
A program is a series of instructions that cause a computer to perform a particular
task.
Actually, a computer program Includes more than instructions; It
also contains the data and memory addresses that the
microprocessor needs to accomplish the task defined by the in1-1

structions. Clearly. if the microprocessor is to perform an addition. it must have two

numbers to add and a destination for the result. The computer program must determine
the sources of the data and the destination of the result as well as specifYing the operation to be performed.
All microprocessors execute instructions sequentially unless one of the instructions
changes the execution sequence or halts the computer (Le.. the processor gets the next
instruction from the next consecutive memory address unless the current instruction
specifically directs It to do otherwise).

Ultimately every program becomes translated into a set of binary numbers. For
example. this is the
program that adds the contents of memory locations
6016 and 6116 and places the result in memory location 6216:

zao

00111010
01100000

00000000
01000111
00111010
01100001
00000000
10000000
00110010
01100010
00000000
This is a machine language. or object. program. If this program
were entered into the memory of a ZBO-based microcomputer. the
microcomputer would be able to execute it directly.

THE PROGRAMMING PROBLEM

There are many difficulties associated with creating programs
as object. or binary machine language. programs. These are
some of the problems:

OBJECT
PROGRAM
MACHINE
LANGUAGE
PROGRAM

1)

The programs are difficult to understand or debug (binary numbers all look the
same. particularly after YOU have looked at them for a few hours).

2)

The programs are slow to enter since you must enter each bit indiVidually.

3)
4)

The programs do not describe the task which YOU want the computer to perform in
anything resembling a human readable format.
The programs are long and tiresome to write.

6)

The programmer often makes careless errors that are very difficult to find.

For example. the following version of the addition object program contains a single
bit error. Try to find it:

00111010
01100000
00000000
01000111
01110010
01100001
00000000
10000000
00110010
01100010
00000000

1-2

Although the computer handles binary numbers with ease. people do not. People find
binary programs long. tiresome. confusing. and meaningless. Eventually. a programmer
may start remembering some of the binary codes. but such effort should be spent more
productively.

USING OCTAL OR HEXADECIMAL

We can improve the situation somewhat by writing instrucOCTAL OR
tions using octal or hexadecimal. rather than binary. numbers.
HEXADECIMAL
We will use hexadecimal numbers In this book because they are
shorter. and because they are the standard for the microprocessor Industry. Table 1-1
defines the hexadecimal digits and their binary eqUivalents. The zao program to add
two numbers now becomes:
3A

60
00
47
3A
61

00
80
32
62

00
At the very least. the hexadeCimal version IS shorter to write and not qUite so tiring to
examine.

Errors are somewhat easier to find in a sequence of hexadecimal digits. The erroneous version of the addition program. in hexadecimal form. becomes:
3A

60
00
47
72
61

00
80
32
62

00
The mistake is easier to spot.
What do we do with this hexadecimal program? The microprocessor understands
only binary instruction codes. The answer is that we must convert the hexadeCimal
numbers to binary numbers. ThiS conversion is a repetitive. tiresome task. People who
attempt It make all sorts of petty mistakes. such as looking at the wrong line. dropping a
bit. or transposing a bit or a digit.
This repetitive. grueling task is. however, a perfect Job for a comHEXADECIMAL
puter. The computer never gets tired or bored and never makes
LOADER
silly mistakes. The idea then is to write a program which takes
hexadecimal numbers and converts them into binary numbers. This is a standard
program provided with many microprocessors; it is called a "hexadecimal loader."
Is a hexadecimal loader worth having? If you are willing to write a program uSing binary
numbers, and you are prepared to enter the program In ItS binary form into the computer. then you will not need the hexadeCimal loader.

1-3

If you choose the hexadecimal loader. you will have to pay a price for It. The hexadeclmalloader is itself a program which you must load Into memory. Furthermore. the
hexadecimal loader will occupy memory - memory that you may want to use in some
other way.
The basIc tradeoff. therefore. IS the cost and memory requirements of the hexadecimal
loader versus the savings In programmer time.
A hexadeCimal loader is well worth its small cost.
A hexadecimal loader certainly does not solve every programmIng problem. The hexadecimal version of the program IS still difficult to read or understand; for example. it
does not distinguish Instructions from data or addresses. nor does the program listing
provide any suggestion as to what the program does. What does 32 or 47 or 3A mean?
Memorizing a card full of codes is hardly an appetizing proposition. Furthermore. the
codes will be entirely different for a different microprocessor. and the program will reqUire a large amount of documentation.
Table 1-1. Hexadecimal Conversion Table
Hexadecimal
Digit

Binary
Equivalent

Decimal
Equivalent

0
1
2
3
4
5

0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111

0
1
2
3
4
5

6
7

8
9
A
B
C

0
E
F

6
7

8
9
10
11
12
13
14
15

INSTRUCTION CODE MNEMONICS

An obvious programming improvement is to assign a name to each instruction
code. The instruction code name is called a "mnemonic", or memory jogger. The
instruction mnemonic should describe in some way what the instruction does.
In fact. every microprocessor manufacturer (they can't remember
PROBLEM
hexadeCimal codes either) provides a set of mnemonics for the
WITH
microprocessor instruction set. You do not have to abide by the
MNEMONICS
manufacturer's mnemonics; there is nothing sacred about them.
However. they are standard for a given microprocessor and therefore understood by all
users. These are the instruction names that you will find In manuals. cards. books. articles. and programs. The problem with selecting instruction mnemonics is that not all instructions have "obvious" names. Some instructions do have obvious names (e.g..
ADD. AND. OR), others have obvious contractions (e.g.. SUB for subtraction. XOR for
exclUSive OR), while still others have neither. The result is such mnemonics as WMP.
PCHL. and even SOB (try and guess what that means!). Most manufacturers come up
with mostly reasonable names and a few hopeless ones. However. users who devise
their own mnemonics rarely seem to do much better than the manufacturer.

1-4

Along with the instruction mnemonics. the manufacturer will usually assign names to
the CPU registers. As with the instruction names. some register names are obvious (e.g..
A for Accumulator) while others may have only historical significance. Again. we will
use the manufacturer's suggestions simply to promote standardization.

zao instruction and register mnemonics. as

zao addition program becomes:

If we use standard
defined by Zilog. our

LD
LD
LD
ADD
LD

A.(60H)
B.A
A.(61H)
A.B
(62Hl.A

r----..,
ASSEMBLY
LANGUAGE
PROGRAM

The program is still far from obvIous. but at least some parts are comprehensible.
ADD A.B is a considerable Improvement over 80: LD does suggest loading data into a
register or memory location. Such a program is an assembly language program.

THE ASSEMBLER PROGRAM

How do we get the assembly language program into the computer? We have to translate It. either into hexadecimal or into binary numbers. You can translate an assembly language program
by hand. instruction by instruction. This IS called hand assembly.
Hand assembly of the addition program's instruction codes may be illustrated as
follows:
Instruction Name
LD
LD
ADD
LD

A.(NNl
B.A
A.B
(NNl.A

Hexadecimal Equivalent
3A

47
80
32

As in the case of hexadecimal to binary conversion. hand assembly IS a rote task which
is uninteresting. repetitive. and sublect to numerous minor errors. Picking the wrong
line. transposing digits. omitting instructions. and misreading the codes are only a few
of the mistakes that you may make. Most microprocessors complicate the task even
further by haVing instructions With different word lengths. Some instructions are one
word long while others are two or three words long. Some instructions require data in
the second and third words: others require memory addresses. register numbers. or
who knows what?

Assembly is another rote task that we can assign to the

microcomputer. The microcomputer never makes any
mistakes when translating codes; it always knows how many
words and what format each instruction requires. The program
that does this job is called an "assembler". The assembler
program translates a user program. or "source" program written with mnemonics. into a machine language program, or
"object" program. which the microcomputer can execute. The
assembler's input is a source program and its output is an object

ASSEMBLER
SOURCE
PROGRAM
OBJECT
PROGRAM

program.

The tradeoffs we discussed in connection with the hexadecimal loader are magnified in the case of the assembler. Assemblers are more expensive. occupy more
memory. and require more peripherals and execution time than do hexadecimal
loaders. While users may (and often dol write their own loaders. few care to write their
own assemblers.

1-5

Assemblers have their own rules that you must learn to abide by. These include the
use of certain markers (such as spaces. commas. semicolons. or colons) In appropriate
places. correct spelling. the proper control information. and perhaps even the correct
placement of names and numbers. These rules typically are a minor hindrance that can
be quickly overcome.

ADDITIONAL FEATURES OF ASSEMBLERS

Earlv assembler programs did little more than translate the mnemonic names of instructions and registers Into their binary equivalents. However. most assemblers now provide such additional features as:
1)
2)

3)
4)
5)
6)
7)

AllOWing the user to assign names to memory locations. Input and output deVices.
and even sequences of Instructions.
Converting data or addresses from various number systems (e.g.. decimal or hexadecimal) to binary and converting characters into their ASCII or EBCDIC binary
codes.
Performing some arithmetic as part.of the assembly process.
Telling the loader program where in memory parts of the program or data should be
placed.
Allowing the user to assign areas of memory as temporary data storage and to
place fixed data in areas of program memOry.
Providing the Information reqUired to include standard programs from program libraries. or programs written at some other time. in the current program.
Allowing the user to control the format of the program listing and the Input and
output devices employed.

All of these features. of course. Involve additional cost and memoCHOOSING

ry. Microcomputers generally have much Simpler assemblers than
AN
do larger computers. but the tendency always is for the size of asASSEMBLER
semblers to increase. You will often have a choice of assemblers.
The Important criterion IS not how many offbeat features the assembler has. but rather
how convenient it is to work With In normal practice.

DISADVANTAGES OF ASSEMBLY LANGUAGE

The assembler. like the hexadecimal loader. does not solve all the problems of
programming. One problem is the tremendous gap between the microcomputer instruction set and the tasks which the microcomputer is to perform. Computer instructions tend to do things like add the contents of two registers. shift the contents of
the Accumulator one bit. or place a new value Into the Program Counter. On the other
hand. a user generally wants a microcomputer to do something like check if an analog
reading has exceeded a threshold. look for and react to a particular command from a
teletypeWriter. or activate a relay at the proper time. An assembly language programmer must translate such tasks Into a sequence of Simple computer instructions. The
translation can be a difficult. time-consuming job.
Fu rthermore. if you are programming in assembly language. you must have detailed
knowledge of the particular microcomputer that you are using. You must know
what registers and instructions the microcomputer has. preciselv how the instructions
affect the various registers. what addreSSing methods the computer uses. and a myriad
of other information. None of this information IS relevant to the task which the
microcomputer must ultimately perform.

In addition. assembly language programs are not portable.

PORTABILITY
Each microcomputer has its own assembly language. which
reflects ItS own architecture. An assembly language program written for the
will

zao

1-6

not run on the Motorola 6800. the Fairchild F8. or the National Semiconductor PACE.
For example. the addition program written for the Motorola 6800 would be:
LDAA
ADDA
STAA

$60
$61
$62

The lack of portability not only means that you won't be able to use your assembly
language program on another microcomputer. but It also means that you won't be able
to use any programs that weren't specifically written for the microcomputer you are
uSing. This is a particular drawback for microcomputers. since these devices are new
and few assembly language programs exist for them. The result. too frequently. is that
you are on your own. If you need a program to perform a particular task. you are not
likely to find it In the small program libraries that most manufacturers provide. Nor are
you likely to find it in an archive. journal article. or someone's old program file. You will
probably have to write It yourself.

HIGH-LEVEL LANGUAGES

The solution to many of the difficulties associated with asJ COMPILER

sembly language programs is to use, instead, "high-level" or
"procedure-oriented" languages. Such languages allow you to describe tasks in
forms that are problem oriented rather than computer oriented. Each statement in
a high-level language performs a recognizable function; it will generally correspond to many assembly language instructions. A program called a compiler translates the high-level language source program into object code or machine language
instructions.

Many different high-level languages eXist for different types of

, FORTRAN
tasks. If. for example. you can express what you want the computer to do In algebraic notation. you can write your program In FORTRAN (Formula
Translation Language). the oldest and one of the most widely used of the high-level
languages. Now. if you want to add two numbers. you just tell the computer:
SUM = NUMB1+NUMB2

That is a lot simpler (and a lot shorted than either the equivalent machine language program or the equivalent assembly language program. Other high-level languages include COBOL (for bUSiness applications). PASCAL (another algebraic language). PL!l (a
combination. of FORTRAN. ALGOL. and COBOL). and APL and BASIC (languages that
are popular for time-sharing systems).

ADVANTAGES OF HIGH-LEVEL LANGUAGES

Clearly, high-level languages make programs easier and faster to write. A common
estimate is that a programmer can write a program about ten times as fast in a
high-level language as compared to assembly language. That IS just writing the program: it does not include problem definition. program deSign. debugging. testing. or
documentation. all of which become simpler and faster. The high-level language program IS. for instance. partly self-documenting. Even if you do not know FORTRAN. you
probably could tell what the statement illustrated above does.
High-level languages solve many other problems associMACHINE
ated with assembly language programming. The high-level
INDEPENDENCE
language has its own syntax (usually defined by a national or
OF HIGH-LEVEL
International standard). The language does not mention the inLANGUAGES
struction set. registers. or other features of a particular computer. The compiler takes care of all such details. Programmers can concentrate on their
own tasks: they do not need a detailed understanding of the underlYing CPU architecture - for that matter. they do not need to know anything about the computer they are
programming.

1-7

Programs written in a high-level language are portableat least. in theory. They will run on any computer or
microcomputer that has a standard compiler for that language.

PORTABILITY
OF HIGH-LEVEL
LANGUAGES

At the same time. all previous programs written in a high-level

language for prior computers are available to you when programming a new computer.
ThiS can mean thousands of programs In the case of a common language like FORTRAN
or BASIC.

DISADVANTAGES OF HIGH-LEVEL LANGUAGES

Well. if all the good things we have said about high-level languages are true. if you
can write programs faster and make them portable besides. why bother with assembly languages? Who wants to worry about registers. instruction codes.
mnemonics. and all that garbage I As usual. there are disadvantages that balance
the advantages.
One obvious problem IS that you have to learn the "rules" or
SYNTAX OF
HIGH-LEVEL
"syntax" of any high-level language you want to use. A hlghlevel language has a fairly complicated set of rules. You will find
LANGUAGES
that it takes a lot of time lust to get a program that is syntactically
correct (and even then It probably will not do what you want). A high-level computer
language is like a foreign language. If you have a little talent. you will get used to the
rules and be able to turn out programs that the compiler will accept. Still. learning the
rules and trying to get the program accepted by the compiler doesn't contribute
directly to dOing your lob.
Here. for example. are some FORTRAN rules:
Labels must be numbers placed In the first five card columns
Statements must start In column seven
Integer variables must start with the letters I. J. K. L. M. or N
Another obVIOUS problem IS that you need a compiler to translate programs written in a high-level language. Compilers are
expensive and use a large amount of memory. While most assemblers occupy 2K to 16K bytes of memory (1 K = 1024), compilers occupy 4K to 64K
bytes. So the amount of overhead Involved in uSing the compiler is rather large.
Furthermore. only some compilers will make the implementation of your task simpler. FORTRAN. for example. IS well-SUited
to problems that can be expressed as algebraic formulas. If.
however. your problem IS controlling a printer. editing a string of characters. or monitorIng an alarm system. your problem cannot be easily expressed In algebraic notation. In
fact. formulating the solution In algebraiC notation may be more awkward and more
difficult than formulating It in assembly language. One answer IS to use a more sUitable
high-level language. Some such languages eXIst. but they are far less Widely used and
standardized than FORTRAN. You will not get many of the advantages of high-level
languages if you use these so-called system Implementation languages.

,..;;..-----

High-level languages do not produce very efficient

INEFFICIENCY
machine language programs. The baSIC reason for this is that
OF HIGH-LEVEL
compilation IS an automatiC process which IS riddled with comLANGUAGES
promises to allow for many ranges of possibilities. The compiler works much like a computerized language translator - sometimes the words are
right but the sounds and sentence structures are awkward. A Simple compiler cannot
know when a variable IS no longer being used and can be discarded. or when a register
should be used rather than a memory location. or when variables have Simple relationships. The experienced programmer can take advantage of shortcuts to shorten execu-

1-8

tlon time or reduce memory usage. A few compilers (known as optimizing compilers)
can also do this. but such compilers are much larger and slower than regular compilers.
The general advantages and disadvantages of high-level languages are:
Advantages:

More convenient descriptions of tasks

More efficient program coding
Easier documentation
Standard syntax
Independence of the structure of a particular computer
Portability
Availability of library and other programs

ADVANTAGES
OF
HIGH-LEVEL
LANGUAGES

Disadvantages:

Special rules
DISADVANTAGES
OF
Extensive hardware and software support required
HIGH-LEVEL
Orientation of common languages to algebraic or business
LANGUAGES
problems
Inefficient programs
Difficulty of optimizing code to meet time and memory requirements
Inability to use special features of a computer conveniently

HIGH-LEVEL LANGUAGES FOR MICROPROCESSORS

Microprocessor users will encounter several special difficulties when using highlevel languages. Among these are:

Few high-level languages eXist for microprocessors

No standard languages are Widely available
Few compilers actually run on microcomputers. Those that do often require very large
amounts of memory.
Most microprocessor applications are not well-suited to high-level languages.
Memory costs are often critical in microprocessor applications.
The lack of high-level languages is partly a result of the fact that microprocessors are
quite new and are the products of semiconductor manufacturers rather than computer
manufacturers.
Very few high-level languages' eXist for microprocessors. The most common are the
PL!1 type languages (such as Intel's PL!M. Motorola's MPL. and Signetlcs' PL/.LSl.
BASIC. and PASCAL.
Even the few high-level languages that exist do not conform to recognized standards.
so the microprocessor user cannot expect to gain much program portability. access to
program libraries. or use of previous experience or programs. The main advantages remaining are the reduction In programming effort and the smaller amount of detailed
understanding of the computer architecture that is necessary.
OVERHEAD
The overhead involved In using a high-level language with
FOR
microprocessors is considerable. Microprocessors themselves are
HIGH-LEVEL
better suited to control and slow interactive applications than they
LANGUAGES
are to the character manipulation and language analysIs involved
In compilation. Therefore. most compilers for microprocessors will
not run on a microprocessor-based system. Instead. they require a much larger computer. I.e.. they are cross-compilers rather than self-compilers. A user must nof only

1-9

bear the expense of the larger computer but must also physically transfer the program
from the larger computer to the micro.
A few self-compilers are available. These compilers run on the microcomputer for
which they produce object code. Unfortunately, they require large amounts of memory
(16K or more). plus special supporting hardware and software.
UNSUITABILITY
High-level languages also are not generally well-suited to
OF HIGH-LEVEL
microprocessor applications. Most of the common languages
LANGUAGES
were devised either to help solve scientific problems or to handle large-scale business data processing. Few microprocessor
applications fall in either of these areas. Most microprocessor applications Involve sending data and control information to output deVices and receiving data and status information from Input deVices. Often the control and status information consists of a few
binary digits with very precise hardware-related meanings. If you try to write a tYPical
control program In a high-level language, you often feel like someone who is trying to
eat soup with chopsticks. For tasks in such areas as test equipment. terminals. navigation systems, signal processing, and bUSiness equipment. the high-level languages
work much better than they do In instrumentation, commUniCations, peripherals, and
automotive applications.
APPLICATION
Applications better suited to high-level languages are those which
AREAS FOR
require large memories. If. as In a valve controller, electronic game,
LANGUAGE
appliance controller. or small Instrument. the cost of a single
LEVELS
memory chip IS important. then the Inefficiency of high-level
languages IS Intolerable. If. on the other hand, as In a terminal or
test equipment, the system has many thousands of bytes of memory anyway. the inefficiency of high-level languages IS not as important. Clearly the size of the program and
the volume of the product are important factors as well. A large program will greatly increase the advantages of high-level languages. On the other hand, a high-volume application will mean that fixed software development costs are not as important as
memory costs that are part of each system.

WHICH LEVEL SHOULD YOU USE?

That depends on your particular application. Let us briefly note some of the factors
which may favor particular levels:
Machine Language:
Virtually no one programs in machine language. Its use cannot be justified considering the low cost of an assembler and
the increase in programming speed an assembler provides.

APPLICATIONS
FOR MACHINE
LANGUAGE

Assembly Language:
Short to moderate sized programs
Applications where memory cost is a factor
Real-time control applications

APPLICATIONS
FOR ASSEMBLY
LANGUAGE

Limited data processing

High-volume applications
More input/output or control than computation
High-Level Languages:
Long programs
Low-volume applications requiring long programs
Applications requiring large memories

1-10

APPLICATIONS
FOR HIGH-LEVEL
LANGUAGE

 More computation than input/output or control

Compatibility with similar applications using larger computers
Availability of specific programs in a high-level language which can be used in
the application
Many other factors are also important. such as the availability of a larger computer for
use in development. experience with particular languages. and compatibility with other
applications.
If hardware will ultimately be the largest cost in your application. or if speed IS critical
you should favor assembly language. But be prepared to spend extra time In software
development In exchange for lower memory costs and higher execution speeds. If software will be the largest cost in your application. you should favor a high-level language.
But be prepared to spend the extra money required for the supporting hardware and
software.
Of course. no one except some theorists will object if you use both assembly and hlghlevel languages. You can write the program originally in a high-level language and then
patch some sections In assembly language. However. most users prefer not to do this
because of the havoc It creates In debugging. testing. and documentation.

HOW ABOUT THE FUTURE?

We expect that the future will tend to favor high-level languages for the following
reasons:

Programs always seem to add extra features and grow larger

Hardware and memory are becoming less expensive
Software and programmers are becoming more expensive
Memory chips are becoming available In larger sizes. at lower
"per bit" cost. so actual savings in chips are less likely

FUTURE TRENDS
IN LANGUAGE
LEVELS

More compilers are becoming available

More SUitable and more efficient high-level languages are being developed
More standardization of high-level languages will occur
Assembly language programming of microprocessors will not be a dYing art any more
than it IS now for large computers. But longer programs. cheaper memory. and more expensive programmers will make software costs a larger part of most applications. The
edge in many applications will therefore go to high-level languages.

WHY THIS BOOK?

If the future would seem to favor high-level languages. why have a book on assembly language programming? The reasons are:
1)
2)
3)
4)
5)

Most current microcomputer users program in assembly language (almost twothirds. according to one recent survey).
Many microcomputer users will continue to program in assembly language since
they need the detailed control that It provides.
No suitable high-level language has yet become widely available or standardized.
Many applications require the efficiency of assembly language.
An understanding of assembly language can help In evaluating high-level
languages.

The rest of this book will deal exclusively with assemblers and assembly language programming. However. we do want readers to know that assembly language is not the
only alternative. You should watch for new developments that may significantly reduce
programming costs if such costs are a major factor in your application.

1-11

REfERENCES
Some overall comparisons of the time required to write various types of programs at
different language levels are in M.H. Halstead. Elements of Software SCience. American
Elsevier. New York. 1977 and In V. Schneider. "Prediction of Software Effort and Project
Duration - Four New Formulas" SIGPLAN Notices. June 1978. pp. 49-55.

1-12

Chapter 2
ASSEMBLERS
This chapter discusses the functions performed by assemblers, beginning with features
common to most assemblers, and proceeding through more elaborate capabilities such
as macros and conditional assemblY, You may wish to skim this chapter for the present
and return to it when you feel more comfortable With the material.

FEATURES OF ASSEMBLERS
As we mentioned previously, today's assemblers do much more than translate assembly language mnemonics into binary codes. But we will first describe how an
assembler handles the translation of mnemonics before describing additional assembler features. Finally, we will explain how assemblers are used.

ASSEMBLER INSTRUCTIONS
Assembly language instructions (or "statements") are divided
into a number of fields, as shown in Table 2-1.

ASSEMBLY
LANGUAGE
FIELDS

The operation code field is the only field which can never be
empty; it always contains either an instruction mnemonic or a
directive to the assembler, called a pseudo-instruction, pseudo-operation, or
pseudo-op.
The address field may contain an address or data, or it may be blank.
Table 2-1. The Fields of an Assembly Language Instruction

Label
Field

Operation
Code or
Mnemonic
Field

Operand
or
Address
Field

START

LD
LD
LD
ADD
LD

A,(VALll
B,A
A,(VAL2l
A,B
(SUM),A

NEXT

VALl'
VAL2:
SUM:

DEFS
DEFS
DEFS

1
1
1

Comment Field

:LOAD FIRST NUMBER INTO A

:SAVE IN B
:LOAD SECOND NUMBER INTO A
:ADD FIRST NUMBER TO A
:STORE SUM
:NEXT INSTRUCTION

The comment and label fields are optional. A programmer will assign a label to a
statement or add a comment as a personal convenience, e.g., to make the program
easier to code and read.

2-1

Of course, the assembler must have some way of telling

FORMAT
where one field ends and another begins. Assemblers that use
punched card Input often require that each field start in a specific card column. This is
a fixed format. However. fixed formats may be inconvenient when the input medium IS
paper tape: fixed formats are also a nuisance to programmers. The alternative is a free
format. where the fields may appear anywhere on the line.

If the assembler cannot use the position In the line to tell the fields
DELIMITERS
apart. it must use something else. Most assemblers use a
special symbol or delimiter at the beginning or end of each field. The most obVIOUS
delimiter IS the space character. Commas. periods. semicolons. colons. slashes. question marks and other characters that would not otherwise be used In assembly
language programs also may serve as delimiters. Table 2-2 lists standard Zilog
assembler delimiters.

zao

Table 2-2. Standard

zao Assembler Delimiters

after a label
'space' between operation code and address
between operands in the address field
before a comment

You will have to exercise a little care with delimiters. Some assemblers are fussy
about extra spaces or the appearance of delimiters in comments or labels. A wellwritten assembler will handle these minor problems, but many assemblers are not
well-written. Our recommendation is simple: avoid potential problems if you can.
The following rules will help:
1)

Do not use extra spaces. particularly after commas that separate operands.

2)

Do not use delimiter characters in names or labels.

3)

Include standard delimiters even if your assembler does not reqUire them. Your programs will then be assembled by any assembler.

LABELS
The label field is the first field in an assembly language instruction: It may be blank. If a label IS present. the assembler
assigns to the label the value of the address for the memory location Into which the first object program byte for that instruction IS loaded. You may
subsequently use the label as data or as an address in another instruction's operand
field. The assembler will replace the label With the assigned value When creating an oblect program.
Labels are most frequently used in Jump, Call or Branch inLABELS
structions. These instructions place a new value In the Program
IN JUMP
Counter and so alter the normal sequential execution of instrucINSTRUCTIONS
tions. JUMP 15016 means "place the value 15016 into the Program Counter" The next instruction to be executed will be the one in memory location
15016 The instruction JUMP START means "place the value assigned to 'the label
START Into the Program Counter" The next Instruction to be executed will be the one
In the memory location to which the label START has been assigned. Table 2-3 contains
an example.

2-2

Table 2-3. Assigning and Using a Label

ASSEMBLY LANGUAGE PROGRAM
START

LOAD ACCUMULATOR 100

(MAIN PROGRAM)

JUMP START
When the machine language version of this program is executed. the instruction JUMP
START causes the address of the instruction labeled START to be placed Into the Program Counter. The instruction with the label START will be executed next.
Why use a label? Here are some reasons:
1)

A label makes a program location easier to find and remember.

2)

The label can be moved to change or correct a program. You do not have to change
any subsequent Instructions that use the label; the assembler will make all the
necessary changes.
The assembler or loader can relocate the whole program by
RELOCATION
adding a constant (a relocation constant) to each address in
CONSTANT
which a label was used. Thus w~ove the program to
allow for the insertion of other programs or simply to rearrange memory.
The program is easier to use as a library program. i.e.. It IS easier for someone else to
take your program and add it to some totally different program.
You do not have to figure out memory addresses. Figuring out memory addresses is
particularly difficult with microprocessors which have Instructions that vary In
length.

3)

4)
5)

It makes sense to assign a label to any instruction that you might want to use as a
destination or otherwise identify.
The next question is what label to use. The assembler often
places some restrictions on the number of characters (usuallv 5
or 6). the leading character (often must be a letter). and the trailing
characters (often must be letters. numbers. or one of a few special characters). Beyond
these restrictions. the choice is up to you.

Our own preference is to use labels that suggest their purpose. I.e.. mnemonic labels.
Typical examples are ADDW in a routine that adds one word Into a sum. SRETX in a
routine that searches for the ASCII character ETX. or NKEYS for a location In data
memory that contains the number of key entries. Meaningful labels are easier to
remember and contribute to program documentation. Some programmers prefer to use
a standard format for labels. such as starting with LOOOO. These labels are self-sequencing (you can skip a few numbers to permit Insertions). but they do not help document
the program.
Some label selection rules will keep you out of trouble. We
recommend the following:

1)

Do not use labels that are the same as operation codes or

other mnemonics. Most assemblers will not allow this usage; others will. but it IS
very confusing.

2-3

2)
3)

Do not use labels that are longer than the assembler permits. Assemblers have
various truncation rules.
Avoid special characters (non-alphabetic and non-numeric) and lower-case letters.
Some assemblers will not permit them: others allow only certain ones. The simplest
practice IS to stick to capital letters and numbers.

4)

Start each label with a letter. Such labels are always acceptable.

5)

Do not use labels that could be confused with each other. Avoid the letters 1.0 and

Z and the numbers O. 1 and 2. Also avoid things like XXXX and XXXXX. There's no
sense tempting fate and Murphy's laws.
6)

When you are not sure if a label IS legal. do not use It. You will not get any real
benefit from discovering exactly what the assembler will accept.

These are recommendations. not rules. You do not have to follow them. but don't blame
us if you waste time on silly problems.

ASSEMBLER OPERATION CODES (MNEMONICS)

The main task of the assembler is the translation of mnemonic operation codes
into their binary equivalents. The assembler performs thiS task using a fixed table
much as you would if you were doing the assembly by hand.
The assembler must. however. do more than lust translate the operation codes. It must
also somehow determine how many operands the instruction requires and what
type they are. This may be rather complex - some instructions (ljke a Halt) have no
operands. others (like an Addition or a Jump Instruction) have one. while still others
(like a transfer between registers or a multiple-bit shift) require two. Some Instructions
may even allow alternatives. e.g.. some computers have Instructions (like Shift or Clearl
that can apply either to the Accumulator or to a memory location. We will not discuss
how the assembler makes these distinctions: we will lust note that it must do so.

PSEUDO-OPERATIONS
Some a,ssembly language instructions are not directly translated into machine language instructions. These instructions
are directives to the assembler; they assign the program to certain areas In memory. define symbols. designate areas of RAM for
temporary data storage. place tables or other fixed data in memory. allow references to other programs. and perform minor housekeeping fu nctions.

PSEUDOOPERATIONS
ASSEMBLER
DIRECTIVE

To use these assembler directives. or pseudo-operations. a programmer places the

pseudo-operation's mnemonic in the operation code field and. if the specified pseudooperation requires It. an address or data in the address field.
The most common pseudo-operations are:
DATA
EQUATE or DEFINE
ORIGIN
RESERVE
Linking pseudo-operations are:
ENTRY
EXTERNAL

2-4

Different assemblers use different names for these operations. but the purposes are the
same. Housekeepmg pseudo-operations mclude:
END
LIST
NAME
PAGE
SPACE
TITLE
We will discuss these pseudo-operations bnefly. although their functions are usually
obvious.

THE DATA PSEUDO-OPERATION

The DATA pseudo-operation allows the programmer to enter fixed data into
memory. This data may mclude:
Lookup tables
Code conversion tables
Messages
Synchronization patterns
Thresholds
Names
Coefficients for equations
Commands
Conversion factors
Weighting factors
Characteristic times or frequencies
Subroutme addresses
Key Identifications
Test patterns
Character generation patterns
Identification patterns
Tax tables
Standard forms
Masking patterns
State transition tables
The DATA pseudo-operation treats the data as a permanent part of the program.

The format of a DATA pseudo-operation is usually quite simple. An instruction

like:
DzeON

DATA

12

will place the number 12 in the next available memory location and assign that
location the name DZeON. Usually every DATA pseudo-operation has a label. unless It
IS one of a series of DATA pseudo-operations. The data and label may take any form
that the assembler permits.
Most assemblers allow more elaborate DATA instructions that handle a large amount of
data at one time. e.g..
EMESS

SQRS

DATA
DATA

2-5

'ERROR'
1,4.9.16.25

A single Instruction may fill many words of program memory. limited only by the length
of a line. Note that if you cannot get all the data on one line. you can always follow one
DATA instruction with another. e.g..
MESSG

DATA
DATA
DATA
DATA
DATA
DATA

'NOW IS THE'
'TIME FOR ALL'
'GOOD MEN'
'TO COME TO THE'
'AID OF THEIR'
'COUNTRY'

Microprocessor assemblers typically have some variations of standard DATA

pseudo-operations. DEFINE BYTE or FORM CONSTANT BYTE handles 8-blt numbers;
DEFINE WORD or FORM CONSTANT WORD handles 16-blt numbers or addresses.
Other special pseudo-operations may handle character-coded data.

THE EQUATE (or DEFINE) PSEUDO-OPERATION

The EQUATE pseudo-operation allows the programmer to
equate labels and names with addresses or data. This pseudooperation is almost always given the mnemonic EQU. The
names may refer to device addresses. numeric data. starting addresses. fixed addresses. etc.
The EQUATE pseudo-operation assigns the numeric value in its operand field to
the label in its label field. Here are two examples:

TTY
LAST

EQU
EQU

5
5000

Most assemblers will allow you to define one label In terms of another. e.g..
LAST
ST1

EQU

EQU

FINAL
START+l

The label in the operand field must. of course, have been previously defined. Often. the
operand field may contain more complex expressions, as we shall see later. Double
name assignments (two names for the same data or address) may be useful In patching
together programs which use different names for the same variable (or different spellIngs of what was supposed to be the same name).
Note that an EQU pseudo-operation does not cause the assembler to place anything into memory. The assembler simply
enters an additional name into a table (called a symbol table)
which the assembler maintains. This table. unlike the mnemonic table, must be In
RAM since It vanes with each program. The assembler program will always need some
RAM to hold the symbol table; the more RAM it has. the more symbols it can accept.
This RAM is In addition to any which the assembler needs as temporary storage.

When do you use a name? The answer IS: whenever you have a
parameter that has some meaning besides ItS ordinary numenc
value. or the numeric value of the parameter might be changed.
We typically assign names to time constants. device addresses. masking patterns. conversion factors, and the like. A name like DELAY. TTY, KBD. NROW. or OPEN not only
makes the parameter easier to change, but It also adds to program documentation. We
also assign names to memory locations that have special purposes; they may hold data.
mark the start of the program. or be available for intermediate storage.
What name do you use? The best rules are much the same as
in the case of labels, except that here meaningful names really
count. Why not call the teletypewnter TTY instead of X15. a bit
time delay BTIME or BTDLY rather than WW. the number of the

2-6

CHOICE
OF
NAMES

"GO" key on a keyboard GOKEY rather than HORSE? This advice seems straightforward. but a surprising number of programmers do not follow It.

Where do you place the EQUATE pseudo-operations? The

PLACEMENT
best place is at the start of the program, under appropriate
OF
comment headings such as I/O ADDRESSES. TEMPORARY
DEFINITIONS
STORAGE, TIME CONSTANTS. or PROGRAM LOCATIONS. This
makes the definitions easy to find if you want to change them. Furthermore. another
user will be able to look up all the definitions In one centralized place. Clearly this practice improves documentation and makes the program easier to use.
Definitions used only in a specific subroutine should appear at the start of the
subroutine.

THE ORIGIN PSEUDO-OPERATION

The ORIGIN pseudo-operation (almost always abbreviated ORG) allows the programmer to locate programs, subroutines, or data anywhere in memory. Programs
and data may be located In different areas of memory depending on the memory configuration. Startup routines. interrupt service routines. and other required programs
may be scattered around memory at fixed or convenient addresses.
The assembler maintains a Location Counter (comparable to
the computer's Program Counter) which contains the location
in memory at which the next byte of object code generated by
the assembler will reside when the program is loaded. An ORG pseudo-operation
causes the assembler to place a new value Into the Location Counter. much as a Jump
Instruction causes the CPU to place a new value Into the Program Counter. The output
from the assembler must not only contain instructions and data. but must also Indicate
to the loader program where In memory it should place the instructions and data.
Microprocessor programs often contain several ORIGIN statements for the following
purposes:
Reset (startup) address
Interrupt service addresses
Trap addresses
RAM storage
Memory stack

Main program
Subroutines
Memory addresses for
Input/output deVices
or speCial functions

Still other ORIGIN statements may allow room for later Insertions. place tables or data in
memory. or assign vacant RAM space for data buffers. Program and data memory In
microcomputers may occupy widely scattered addresses to simplify the hardware.

Typical ORIGIN statements are:

ORG
ORG
ORG

RESET
1000
INT3

Some assemblers assume an Origin of zero if the programmer does not put an ORG
statement at the start of the program. The convenience is slight: we recommend the inclUSion of an ORG statement to avoid confusion.

THE RESERVE PSEUDO-OPERATION

The RESERVE pseudo-operation allows the programmer to
allocate RAM for various purposes such as data tables, temporary storage, indirect addresses, a Stack, etc.

2-7

ALLOCATING

RAM

Using the RESERVE pseudo-operation, you assign a name to the memory area and
declare the number of locations to be assigned. Here are some examples:
NOKEY
TEMP
VOLTG
BUFR

RESERVE
RESERVE
RESERVE
RESERVE

1
50

80
100

You can use the RESERVE pseudo-operation to reserve memory locations in program
memory or in data memory: however the nature of the RESERVE pseudo-operation is
more meaningful when applied to data memory.
In reality, all the RESERVE pseudo-operation does is Increase the assembler's Location
Counter by the amount declared In the operand field. The assembler does not actually
produce any oblect code.

Note the following features of RESERVE:

1)

The label of the RESERVE pseudo-operation IS assigned the value of the first address reserved. For example, the sequence:
BUFl
BUF2
VOLTS

2)
3)

ORG
RESERVE
RESERVE
RESERVE

3000
100
50

assigns to the label BUFl the value 3000, to BUF2 3100, and to VOLTS 3150.
You must specify the number of locations to be reserved. There IS no default case.
No data IS placed into the reserved locations. Any data that by chance, may be In
these locations will be left there.

Some assemblers allow the programmer to place initial

INITIALIZING
values in RAM. We strongly recommend that you do not
RAM
use this feature - it assumes that the program (along with
the initial values) will be loaded from an external deVice (e.g., paper tape or floppy disk)
each time it IS run. Most microprocessor programs, on the other hand, reSide In nonvolatile ROM and start when power comes on. The RAM In such situations does not retain its contents, nor IS It reloaded. Always include instructions to Initialize the RAM In
your program.

LINKING PSEUDO-OPERATIONS
We often want statements in one program or subroutine to
EXTERNAL
use names that are defined elsewhere. Such names are called
REFERENCES
external references; a special linker program IS necessary to actually fill in the external values and determine if any names are undefined or doubly
defined.
The pseudo-operation EXTERNAL, usually abbreviated EXT, signifies that the
name is defined elsewhere.
The pseudo-operation ENTRY, usually abbreviated ENT, signifies that the name is
available for use elsewhere, I.e., It is defined In thiS program.
The precise way in which linking pseudo-operations are Implemented vanes greatly
from assembler to assembler. We will not refer to such pseudo-operations again, but
they are very useful in actual applications.

HOUSEKEEPING PSEUDO-OPERATIONS
There are various housekeeping pseudo-operations, which affect the operation of

2-8

the assembler and its program listing rather than the output program itself. Common housekeeping pseudo-operations include:
1)

END. which marks the end of the assembly language source program.

2)

LIST. which tells the assembler to print the source program. Some assemblers allow
such variations as NO LIST or LIST SYMBOL TABLE to avoid long. repetitive listIngs.

3)

NAME or TITLE. which prints a name at the top of each page of the listing.

4)

P.AGE or SPACE. which skips to the next page or next line. respectively. and Improves the appearance of the listing. making It easier to read.

5)

PUNCH. which transfers subsequent object code to the paper tape punch. This
pseudo-operation may In some cases be the default option and therefore unnecessary.

LABelS WITH PSEUDO-OPERATIONS

Users often wonder if or when they can assign a label to a pseudo-operation.
These are our recommendations:
1)

All EQUATE pseudo-operations must have labels: they do not make any sense
otherwise. since their purpose IS to define the meaning of the labels.

2)

DATA and RESERVE pseudo-operations usually have labels. The label identifies the
first memory location used or assigned.

3)

Other pseudo-operations should not have labels. Some assemblers allow other
pseudo-operations to have labels. but the meaning of the labels vanes. We recommend that you aVOid thiS practice.

ADDRESSES AND THE OPERAND FIELD

Most assemblers allow the programmer a lot of freedom in describing the contents of the Operand Address field. But remember, the assembler has built-in
names for registers and instructions and may have other built-in ...
na,;;m~e_s;..' _ _.....
Some common options for the operand field are:

DECIMAL

1)

DATA OR
ADDRESSES

Decimal numbers

Most assemblers assume all numbers to be deCimal unless they

are marked otherwise. So:

100

ADD

means "add the contents of memory location 100 deCimal to the contents of the Accumulator"

2)

Other number systems

Most assemblers will also accept binary. octal. or hexadeCimal entries. But you must identify these number systems In some way.
e.g.. by preceding or follOWing the number with an Identifying
character or letter. Here are some common Identifiers:

OTHER
NUMBER
SYSTEMS

B or % for binary
O. Q. C or @ for octal (we aVOid 0 because of the confusion with zero).
H or $ for hexadeCimal

o for decimal. D may be omitted:

it is the default case.

2-9

Assemblers generally require hexadecimal numbers to start with a decimal digit (e.g..
OA36 Instead of A36) In order to distinguish between numbers and names or labels. It is
good practice to enter numbers in the base in which their meaning is the clearestI.e.. decimal constants in deCimal: addresses and BCD numbers In hexadecimal: maskIng patterns or bit outputs In binary if they are short and in hexadecimal if they are long.

3)

Symbolic names

Names can appear in the operand field: they will be treated as the data that they represent. But remember. there is a difference between data and addresses. The sequence:

FIVE

EQU
ADD

5
FIVE

will add the contents of memory location 5 (not necessarily the number 5) to the contents of the Accumulator.

4)

The current value of the location counter (usually referred to as or $).

This IS useful mainly In Jump instructions: for example:

JUMP

$+6

causes a Jump to the memory localion six words beyond the word that contains the
first byte of the JUMP instruction:
Memory

t----il}

JUMP

+6

code stored here

6 locations

Jump here

Most microprocessors have many two and three-word Instructions. Thus. you will have
difficulty determining exactly how far apart two assembly language statements are.
Therefore. using offsets from the Location Counter frequently results in errors that you
can avoid if you use labels.

5)

Character codes

Most assemblers allow text to be entered as .ASCII strings, Such

strings may be surrounded either With Single or double quotation
marks: strings may also use a beginning or ending symbol such as
A or C, A few assemblers also permit EBCDIC strings,

ASCII
CHARACTERS

We recommend that you use character strings for all text. It Improves the clarity and
readability of the program.

6)

Combinations of 1) through 5) with arithmetic, logical, or special operators.

Almost all assemblers allow simple arithmetic combinations such

ARITHMETIC
as START+l, Some assemblers also permit multiplication. divlAND LOGICAL
sion. logical functions. shifts. etc, These are referred to as expresEXPRESSIONS
sions, Note that the assembler evaluates expressions at assembly
time. Even though an expression In the operand field may Involve multiplication. you

2-10

may not be able to use multiplication In the logic of your own program write a subroutine for that specific purpose.

unless YOU

Assemblers vary In what expressions they accept and how they Interpret them. Complex expressions make a program difficult to read and understand.
We have made some recommendations dUring this section but will repeat them and
add others here. In general, the user should emphasize clarity and simplicity. There
IS no payoff for being an expert in the intricacies of assemblers or in having the most
complex expression on the block. We suggest the following approach:
1)

2)
3)
4)

Use the clearest number system or character code for data. Masks and BCD numbers in decimal. ASCII characters in octal. or ordinary numerical constants In hexadecimal serve no purpose and therefore should not be used.
Remember to distinguish data and addresses.
Don't use offsets from the Location Counter.
Keep expressions simple and obvious. Don't rely on obscure features of the assembler.

CONDITIONAL ASSEMBLY
Some assemblers allow you to include or exclude parts of the source program. depending on conditions existing at assembly time. This is called conditional assembly; It gives the assembler some of the flexibility of a compiler. Most microcomputer
assemblers have limited capabilities for conditional assembly. A usual form is:
IF COND
.CONDITIONAL PROGRAM

ENDIF
If the expression COND IS true at assembly time, the instructions between IF and ENDIF
(two pseudo-operations) are Included in the program.
Typical uses of conditional assembly are:
1)
2)
3)
4)

To
To
To
To

include or exclude extra variables.

place diagnostics or special conditions in test runs.
allow data of various bit lengths.
create specialized versions of a common program.

Unfortunately, conditional assembly tends to clutter programs and make them difficult
to read. Use conditional assembly only if It is necessary.

MACROS
You will often find that particular sequences of instructions ocDEFINING A
cur many times in a source program. Repeated instruction seSEQUENCE OF
INSTRUCTIONS
quences may reflect the needs of your program logic, or they
may be compensating for deficiencies in your microprocessor's
instruction set. You can avoid repeatedly writing out the same instruction sequence by
uSing a macro.
Macros allow you to assign a name to an instruction sequence. You then use the
macro name in your source program instead of the repeated instruction sequence.

2-11

The assembler will replace the macro name with the, appropriate sequence of instructions. This may be illustrated as follows:

Object Program

Source Program
MAC'

(macro definitloni

MACRO
instruction M 1 }
instruction M2

---------------.

instruction M3

ENDM

(end of macro definition!

instructIon P 1

(ma,n programl }

-+

...

instructIon P2
Instruction

P3

instruction Pl
Instruction P2
Instruction P3

Instruction

------------------t'''----"I

MAC'

instruction P4

instruction

--------------+---t~

P7

instruction

instruction P9

MAC1 - - - - - - - - - - - - - - - - - - '......

instruction P11

... {

l
i

...

instruction

P7

Instruction

M1

instruction M2
Instruction M3

-{
.. {
--1_

instruction PB

P5

Instruction P6

-----------------f""---t....

instruction P10

M1

Instruction M2
instruction M3
instruction P4

tnstrucliOn P5
instruction P6

MAC'

Instruction

pa
P9

Instruction

M1

instruction

instruction M2
instruction

M3

instruction P10
instruction Pl1

Macros are not the same as subroutines. A subroutine occurs once in a program. and
program execution branches to the subroutine. A macro is expanded to an actual instruction sequence each time the macro occurs: thus a macro does not cause any
branching.

Macros have the following advantages:

1)

Shorter source programs.

ADVANTAGES
OF MACROS

2)

Better program documentation.

3)

Use of debugged instruction sequences - once the macro has been debugged.
you are sure of an error-free Instruction sequence every time you use the macro.
EaSier changes. Change the macro definition and the assembler makes the change
for you every time the macro is used.

4)
5)

InclUSion of commands. keywords. or other computer instructions In the basic instruction set. You use the macro as an extension of your Instruction set.

The disadvantages of macros are:

1)

DISADVANTAGES

Repetition of the same instruction sequences since the

macro IS expanded every time It IS used.

2-12

OF MACROS

2)
3)
4)

A single macro may create a lot of instructions.

Lack of standardization that may make the program difficult to read and understand.
Possible effects on registers and flags that may not be clearly stated.

One problem IS that variables used In a macro are known only

within it (i.e.. they are local rather than globall. This can often
create a great deal of confusion without any gain In return. You
should be aware of this problem when using macros.

----..,
LOCAL OR
GLOBAL
VARIABLES

COMMENTS
All assemblers allow you to place comments in a source program. Comments have
no effect on the object code. but they help you to read. understand. and document
the program. Good commenting is an essential part of writing assembly language
programs; without comments. programs are very difficult to understand.
We will discuss commenting along with documentation in a
later chapter. but here are some guidelines:
1)

COMMENTING
TECHNIQUES

Use comments to tell what the program IS doing, not what instructions do,
Comments should say things like "IS TEMPERATURE ABOVE L1MIT?", "LINE FEED
TO TTY". or "EXAMINE LOAD SWITCH"
Comments should not say things like "ADD 1 TO ACCUMULATOR" "JUMP TO
START", or "LOOK AT CARRY" You should describe how the program IS affecting
the system: internal effects on the CPU are seldom of any Interest.

2)

Keep comments brief and to the POint. Details should be available elsewhere in the
documentation,

3)

Comment all key POints.

4)

Do not comment standard instructions or sequences that change counters and

pointers: pay special attention to Instructions that may not have an obvious meanIng,

5)

Do not use obscure abbreviations.

6)

Make the comments neat and readable.

7)

Comment all definitions, describing their purposes. Also mark all tables and data
storage areas.

8)

Comment sections of the program as well as individual instructions.

9)

Be consistent in your terminology, You can (should) be repetitive: you do not need
to consult a thesaurus,

10)

Leave yourself notes at points which you find confusing, e.g.. "REMEMBER CARRY WAS SET BY LAST INSTRUCTION" You may drop these In the final documentation.

A well-commented program is easy to work with. You will recover the time spent In
commenting many times over. We will try to show good commenting style In the programming examples, although we often over-comment for instructional purposes.

2-13

TYPES Of ASSEMBLERS
Although all assemblers perform the same tasks, their implementations vary
greatly. We will not try to describe all the existing types of assemblers; we will
merely define the terms and indicate some of the choices.
A cross-assembler is an assembler that runs on a computer
other than the one for which it assembles object programs.
The computer on which the cross-assembler runs IS typically a
large computer with extensive software support and fast peripherals - such as an IBM
360 or 370. a Univac 1108. or a Burroughs 6700. The computer for which the cross-assembler assembles programs is typically a microcomputer like the Z80 or MC6800.
Most cross-assemblers are written in FORTRAN so that they are portarb..;le;...
.,
A self-assembler or resident assembler is an assembler that runs
on the computer for which It assembles programs. The self-assembler will require some memory and peripherals. and it may run
qUite slowly.

A macroassembler is an assembler that allows you to define

sequences of instructions as macros.

MACROASSEMBLER

A microassembler is an assembler used to write the

microprograms that define the instruction set of a computer.
Microprogramming has nothing specifically to do with
microcomputers.

MICROASSEMBLER

A meta-assembler is an assembler that can handle many

different instruction sets. The user must define the particular instruction set being used.
A one-pass assembler is an assembler that goes through the
assembly language program only once. Such an assembler must
have some way of resolving forward references. e.g.. Jump instructions which use labels that appear later In the source program. I.e.. that have not
yet been defi ned.
A two-pass assembler is an assembler that goes through the
TWO-PASS
assembly language source program twice. The first time the
ASSEMBLER
assembler simply collects and defines all the symbols; the
second time it replaces the references with the actual definitions. A two-pass assembler solves most of the forward reference problems. However, macro expansion and conditional assembly can cause problems. On some large machines seven
or more passes are needed to insure that all forward references are resolvable. A
two-pass assembler may be quite slow if no backup storage !like a floppy diskl is
available; then the assembler must physically read the program twice from a slow
input medium !like a teletypewriter paper tape readerl. Most microprocessorbased assemblers require two passes.

ERRORS
Assemblers normally provide error messages, often consisting of a single coded
letter. Some typical errors are:
1)
2)

Undefined name (often a misspelling or an omitted definition!.

Illegal character (e.g.. a 2 in a binary number!.

2-14

3)
4)
5)
6)
7)

8)
9)

10)

Illegal format (wrong delimiter or Incorrect operands),

Invalid expression (e.g.. two operators in a row!.
Illegal value (usually too large).
Missing operand.
Double definition (Le.. two different values assigned to one name!.
Illegal label (e.g.. a label on a pseudo-operation that cannot have one!.
Missing label.
Undefined operation code.

In interpreting assembler errors. you must remember that the assembler may get off on
the wrong track if It finds a stray letter. an extra space. or Incorrect punctuation. Many
assemblers will then proceed to misinterpret the succeeding instructions and produce
meaningless error messages. Always look at the first error very carefully: subsequent
ones may depend on it. Caution and consistent adherence to standard formats will
eliminate many annoying mistakes.

LOADERS
The loader IS the program which actually takes the output (oblect codel from the assembler and places it in memory. Loaders range from the very Simple to the very complex. We will describe a few different types.

A bootstrap loader is a program that uses its own first few instructions to load the rest of itself or another loader program
into memory. The bootstrap loader may be In ROM. or you may
have to enter it into the computer memory using front panel switches.
may place a bootstrap loader at the start of the obJect program that It

BOOTSTRAP
LOADER
The assembler
produces.

A relocating loader can load programs anywhere in memory. It

RELOCATING
typically loads each program into the memory space Immediately
LOADER
following that used by the prevIous program. The programs.
however. must themselves be capable of being moved around In thiS way. I.e.. they
must be relocatable. An absolute loader. in contrast. will always place the programs in
the same area of memory.
A linking loader loads programs and subroutines that have
been separately assembled; it resolves external references that is. an instruction In one module that refers to a label In
another module. Object programs loaded by a linking loader must be created by an assembler that permits and marks external references.
An alternative approach IS to separate the linking and loading
functions and have the linking performed by a program called a
link editor.

2-15

REFERENCES
A complete monograph on macros IS M. Campbell-Kelly. An Introduction to Macros.
American Elsevier. New York. 1973.
Microprogramming IS described conceptually in An Introduction to Microcomputers: Volume 1 - Basic Concepts. Chapter 4. A more technical description IS in A.K.
Agrawala and T.G. Rauscher. Foundations of Microprogramming. AcademiC Press. New
York. 1976.
You can find more detailed desCriptions of assemblers and loaders In D.W. Barron. "Assemblers and Loaders". American Elsevier. New York. 1972 and In C.W. Gear. Computer Organization and Programming. McGraw-HilI. New York. 1974.

2-16

Chapter 3

THE

zao ASSEMBLY LANGUAGE

INSTRUCTION SET

We are now ready to start writing assembly language programs. We begin in this
chapter by defining the individual instructions of the zao assembly language instruction set, plus the syntax rules of the Zilog assembler.
We do not discuss any aspects of microcomputer hardware, signals. interfaces. or
CPU architecture in this book. This information IS described in detail In An Introduction
to Microcomputers: Volume 2 -Some Real Microprocessors and Volume 3 -Some
Real Support Devices. while
Programming for Logic Design discusses assembly
language as an extension of digital logic. In this book. we look at programming techniques from the assembly language programmer's viewpoint, where pins and signals are irrelevant and there are no important differences between a minicomputer and a microcomputer.

zao

zao

Interrupts. direct memory access. and the Stack architecture for the
will be described In later chapters of this book. In conjunction With assembly language programming discussions of the same subjects.
This chapter contains a detailed definition of each assemblv language instruction.
Programming for
These definitions are Identical to those found In Chapter 6 of
Logic Design.

zao

The detailed description of individual instructions IS preceded by a general discussion

instruction set that divides instructions Into those which are commonly
of the
used. Infrequently used. and rarely used. If you are an experienced assembly language
programmer. this categorization is not particularly Important - and. depending on your
own programming prejudices. It may not even be accurate. If you are a novice assembly
language programmer. we recommend that you begin by writing programs using only
instructions in the "commonly used" category. Once you have mastered the concepts
of assembly language programming. you may examine other instructions and use them
where appropriate.

zao

3-'

CPU REGISTERS AND STATUS FLAGS

The CPU registers and status flags for the l80 may be illustrated as follows:
Sign
Zero
Auxiliary Carry

,...--- Panty I Overflow

I
s

Iz

II
,

lAd

IPlol N

Ic

Accumulators

s', Z',

Alternate {
Secondary
Accumulators

} Secondarv Data Counters

Primary Data Counter

Flags

Accumulator

secondarv{

Subtract

Carry

IP/o,1 N' , C'

lAd

Alternate Flags

A'

Alternate Accumulator

B'

C'

0'

E'

Alternate Secondary
} Data Counters

L'

H'

Alternate Primary Data Counter

Stack Pointer

SP
PC

Program Counter

IX

Index Register
Index Register

IY

Interrupt Vector Register

Refresh Register

The Accumulator is the primary source and destination for one-operand and twooperand instructions. For example, the shortest and fastest data transfers between the
CPU and I/O devices are performed through the Accumulator, In addition, more Memory Reference instructions move data between the Accumulator and memory than between any other register and memory, All 8-blt arithmetic and Boolean instructions take
one of the operands from the Accumulator and return the result to the Accumulator. An
instruction must therefore load the Accumulator before the l80 can perform any 8bit arithmetic or Boolean operations.
The B. C. D. E. H. and L registers are all secondary registers. Data stored in anv of
these SIX registers may be accessed with equal ease; such data can be moved to any
other register or can be used as the second operand In two-operand instructions,
There are, however, some important differences
H, and L.

In

the functions of Registers B, C, D, E,

Registers Hand L are the primary Data Pointer for the l80. That IS to say, you will
normally use these two registers to hold the 16-blt memory address of data being accessed, Data may be transferred between any registers and the memory location addressed by Hand L. Since HL IS the pnmary Data Pointer, it often takes fewer bytes of
oblect code and less instruction cycles to perform operations with it. The Z80 programmer should try to address data memory via Registers Hand L whenever possible,
Within your program logic. always reserve Registers Hand L to hold a data memoryaddress.

3-2

Registers B, C, 0, and E provide secondary data storage; frequently. the second

operand for two-operand instructions is stored In one of these four registers. (The first
operand is stored in the Accumulator. which IS also the destination for the result.)
There are a limited number of instructions that treat Registers Band C, or 0 and E,
as 16-bit Data Pointers. But these Instructions move data between memory and the
Accumulator only.

In your program logic you should normally use Registers B, C, 0, and E as temporary storage for data or addresses.
Registers IX and IV are index registers. They proyide a limited Indexing capability of
the type described in An Introduction to Microcomputers: Volume 1 for short instructions.
The alternate registers f', A', B', C', 0', E', H', and L' provide a duplicate set of
general purpose registers. Just two single-byte Exchange instructions select and
deselect all alternate registers; one Instruction exchanges AF and the alternate AF'
as a register pair. and one instruction exchanges BC. DE. and HL With the alternate BC',
DE'. and HL' Once selected. all subsequent register operations are performed on the active set until the next exchange selects the inactive set. The alternate registers can be
reserved for use when a fast interrupt response is required. Or. they may be used in
any desired way by the programmer.
There are a number of instructions that handle 16 bits of data at a time. These instructions refer to pairs of CPU registers as follows:

F
B
D
H

F'
B'
D'

H'

and
and
and
and
and
and
and
and

A
C
E
L
A'

C'
E'
L'

'-v-'

'-v-'

Highorder
byte

Loworder
byte

The combination of the Accumulator and flags. treated as a 16-blt unit. is used only for
Stack operations and alternate register switches. ArithmetiC operations access Band C.
D and E. or Hand L as 16-bit data units.

The Carry status flag holds carries out of the most significant bit in any arithmetic
operation. The Carry flag IS also Included in Shift instructions: It IS reset by Boolean instructions.
The Subtract flag is designed for internal use during deCimal adjust operations. This
flag is set to 1 for all Subtract instructions and reset to 0 for all Add Instructions.
The Parity/Overflow flag is a multiple use flag, depending on the operation being
performed. for arithmetic operations, it is an overflow flag. for input, rotate, and
Boolean operations, it is a parity flag, with 1 = even parity and 0 = odd parity. During block transfer and search operations. It remains set until the byte counter decrements to zero: then it is reset to zero. It IS also set to the current state of the Interrupt
enable flip-flop (JFF2) when a LD A.l or LD A.R instruction IS executed.
The Zero flag is set to 1 when any arithmetic or Boolean operation generates a
zero result. The Zero status is set to 0 when such an operation generates a nonzero result.

3-3

The Sign status flag acquires the value of the most significant bit of the result
following the execution of any arithmetic or Boolean instruction.
The Auxiliary Carry status flag holds any carry from bit 3 to 4 resulting from the
execution of an arithmetic instruction. The purpose of this status flag IS to simplify
Binary-Coded-Decimal (BCD) operations; this IS the standard use of an Auxiliary Carry
status flag as described In An Introduction to Microcomputers; Volume 1. Chapter 3.
All of the above status flags keep their current value until an instruction that modifies
them IS executed. Merely changing the value of the Accumulator will not necessarily
change the value of the status flags. For example, if the Zero flag IS set. and a load Immediate to the Accumulator is executed, that causes the Accumulator to acquire a nonzero value; the value of the Zero flag remains unchanged.

The 16-bit Stack Pointer allows you to implement a Stack anywhere in addressable memory. The size of the Stack IS limited only by the amount of addressable memory
present. In reality you will rarely use more than 256 bytes of memory for your Stack.
You should use the Stack for accessing subroutines and processing interrupts. Do not
use the Stack to pass parameters to subroutines. This IS not very efficient Within the
limitations of the
instruction set. The
Stack IS started at ItS highest address. A
Push decrements the Stack POinter contents; a Pop Increments the Stack POinter contents.

zao

zao

The Interrupt Vector register and the Refresh register are special-purpose
registers not normally used by the programmer.
The Interrupt Vector register IS used to store the page address of an Interrupt response
routine; the location on the page IS prOVided by the Interrupting device. This scheme
allows the address of the Interrupt response routine to be changed while still providing
a very fast response time for the interrupting device.
The Refresh register contains a memory refresh counter In the low-order seven bits.
This counter IS Incremented automatically after each Instruction fetch and prOVides the
next refresh address for dynamic memories. The high-order bit of the Refresh register
will remain set or reset. depending on how it was loaded at the last LD R,A instruction.

ZSO MEMORY ADDRESSING MODES

The

zao provides extensive addressing modes. These include:

Implied

Implied Block Transfer with Auto-Increment/Decrement

Implied Stack
Indexed
Direct
Program Relative
Base Page
Register Indirect
Immediate

3-4

Implied
In implied memory addressing. the Hand L registers hold the address of the
memory location being accessed. Data may be moved between the Identified memory location and anyone of the seven CPU registers A. B. C. D, E. H, or L. For example, the
instruction
LD

C,(HL)

loads the C register with the contents of the memory location currently pOinted to by
HL. This IS illustrated as follows:
S ZACP/ON

FD

I I I

Data

Cl

Memory

B.C
D,E
H.L

,.

qq

PP

SP
PC

mmmm

~mmmm+'

.......

IX
IY
I
R

I
I

LD

~
76543210

Load Implied

Via

C RegIster

3-5

HL

P q

Program

Memory

~
4E

C. (HLI

L[T

yy

mmm

m
mm mm+l
mmmm+2
mmmm+3

A limited number of instructions use Registers Band C or D and E as the Data

Pointer. These instructions move data between the Accumulator and the memory location addressed by Registers Band C or Registers D and E. The Instruction
LD

(BCl.A

stores the contents of A Into the memory location currently addressed by Register Pair
Be. This is illustrated as follows:
SZACPONC

Data
Memory

I I ! I I f I
A

ppqq

yy
qql

pp

S.C

II

D.E
H.L
SP
PC

... .r

~mmmm+l

mmmm

1
Program

~
Memory

IX

IY
I

LD

02

(BCIA

A ..._

r.,

54

...._
2

...
0-----

"

-./
store Implied from A via

3-6

Be

mm mm
mm mm+l
mmmm+2
mmmm+3

Implied Block Transfer With Auto-Increment/Decrement

Block Transfer and Search instructions operate on a block of data whose size is
set by the programmer as the contents of the BC register pair. In this form of addressing. a byte of data is moved from the memory location addressed by HL to
the memory location addressed by DE; then HL and DE are incremented and BC is
decremented. Data transfer continues until BC reaches zero. at which point the instruction is terminated. Variations include allowing other instructions to follow
each data transfer. with the programmer supplying the loopback; auto-decrementing HL and DE instead of auto-incrementing; and a complementary set of Block
Search instructions that compare the memory byte addressed by HL with the contents of the A register. setting a flag if a match is found.
The Load. Increment. and Repeat instruction
LDIR
IS

illustrated as follows:

Set if BC- 1."& 0 .eset otherwise

S

AcPYON

FO:m{(]IJ
A
B.C

tt

O.E

rr

ss

.L

pp

qq

uu

SP
PC
IX
IV

Program

mmmm

mory

~
ED
80

mmmm
mmmm+ 1
mmmm+2
mmmrri+3

t-++-+'""i~1-"'"+-0+:~}Load. Increment. and Repeat instruction

A similar group of Input/Output instructions is provided. allowing a block of data
to be input or output between memory and an I/O device. The I/O port number IS
taken as the contents of the C register. with the single B register used as the byte
counter. Memory is addressed by HL.

3-7

Implied Stack
Since the Stack is part of ReadlWrite memory, we must consider Stack Instructions as
Memory Reference Instructions. Push and Pop instructions move two bytes of data
between a register pair and the addressed Stack Pointer location, i.e.. current topof-stack. The ZSO Stack address IS decremented with each Push and Incremented with
each Pop. The instruction
PUSH

DE

is illustrated as follows:
S

ZAcP/ON C

no

Data
Memory

A
B.C

D.E

pp

qq

H.L

SP

SSSS

PC

mmmm

IX
IV
I

ssss

Program
Memory

~
05

PUSH

DE

LZ_
76543210

~~I'IOlqOlqOI1_!_ _
-

PUSH Instruction
Register Pair DE

3-8

ssss - 2
ssss - 1

ssss-2 : )

f# ... - ,
~---:I.. ..,:mmm + 1

qq
pp

mmmm

mmmm+l
mmmm+2
mmmm+3

The zao also has instructions that exchange the two top-of-stack bytes with a
16-bit register - HL or one of the two Index registers. The instruction
EX
IS

(SPl.HL

illustrated as follows:
S

FI

Z AC Pia N C

Data
Memory

! I ! ! I I

qq

B.C
D.E

pp

H.L

II'

xx

yy

SP

ssss

PC
IX

mmmm

IY
I

ssss
ssss + 1
ssss + 2

__

-~mmm+:V

Program

Memory

~
E3

3-9

mmmm

mmmm+ 1
mmmm+2
mmmm+3

Indexed
The zao has two 16-bit index registers, called IX and IV. They may be used Interchangeably. All memory reference operations for which (HL) can be specified can
alternatively be specified as an indexed operation. The difference between implied addressing uSing HL and indexed addressing uSing IX and IY is that the index operand
includes a displacement value that is added to the index address. In the instruction
ADD

A.(IX+40H)

the memory address IS the sum of the contents of the IX register and 4016' This may be
illustrated as follows:
S

ZAc;P/ON C

Data

F~
A

xx

-, xx+yy

B.C
D.E

t:=:=:J

ppqq

...

VY

ippqqi

H.L

SP
PC

mmmm

IX
IY
I
R

ppqq

~mmm+.D

I
ADO

-- -

'-

.J-

A.UX + 40)

':T__.._

~,

16543210

1 0
1 1
0 1 } Add to A Indexed by IX Instruction
0000110

Program
Memory
00

~pqq+~

Displacement

3-10

40

8~

40

mmm m
mmmm+l
mmm m+2
mmm m+3

Direct
Direct addressing can be used to load the Accumulator with any 8-bit value from
memory. load Be. DE. HL. SP.IX. or IV with any 16-bitmemory value. and jump or
call subroutines direct at any memory location. The 16-blt direct address is stored in
the last two bytes of the instruction, In low-byte high-byte order (this IS the reverse of
the standard high-low scheme),

The instruction
LD

A,(NETX)

loads the A register with the contents of the memory location addressed by the label
NETX, The instruction
LD

HL.(1 FFH)

loads the L register With the contents of memory location 01FF 16 and the H register
with the contents of memory location 0200 16, ThiS may be illustrated as follows:
S
F

Z AcP/ON C

Data
Memory

I I I

A
S,C

xx

D,E
H,L

yy

SP

"--~mmm+~

mmmm

PC

IX
IV
I
I

I
R

LD

yy

01FF

xx

0200

Program
Memory
mmmm
mmmm+ 1
1
01
mmmm+2
1-"";';"---1 mmmm + 3
2A

FF

HL,( 1FFH)

543

Load HL Direct Instruction

Direct address - low byte

Direct address - High byte

The direct Jump instructions provide jumps and jumps-to-subroutines. both unconditional and conditional. These are all 3-byte instructions, with the direct address
stored in the second and third bytes of the Instruction, as shown above for Load Direct.

There are three additional addreSSing modes used by Z80 8ranch instructions: program relative, base page, and register Indirect. In general, they are shorter
and/or faster than direct Jumps but may have more limited addreSSing capabilities.

3-11

Program Relative
Jump Relative instructions provide program relative addressing in the range -126,
+129 bytes from the first byte of the Program Relative instruction. These instructions
are all 2-byte instructions, with the signed displacement value stored In the second
byte of the instruction. There are unconditional and conditional relative jumps, as
well as a Decrement and Jump If Not Zero instruction (DJNZ) that facilitates loop
control.

Given the instruction

JR

SRCH

assume that SRCH is a label addressing a location 5A

bytes up In memory from the
16
JR op-code byte. The operation may be illustrated as follows:
S

ZACP/ON

Data

FI I ! ! I I I
A

s.c
D,E
H.l

SP
mmmm

PC
IX

~mmmm+
:::;;;a:
5A

IY

I
I

I
R

JR

-'SRCH

~++~-+~t-~+~ Jump Relative instruction

1I-.l-..I.....l....I._I-.l--'-O.... Displacement

3-12

~
Program
Memory
18
5A

mmmm
mmmm+l
mmmm+2
mmmm+3

Base Page
The Z80 has a modified base page addressing mode for the Restart instruction. This IS
a special Call instruction that allows a single-byte instruction to jump to one of
eight subroutines located at specific points in lower core. The effective address IS
calculated from a 3-bit code stored in the instruction. as follows:
Lower Core Address

3-Bit Code

OOH

000
001
010
011
100
101
110
111

08H
10H
18H
20H
28H
30H
38H

The decoded address value is loaded into the low-order bvte of the Program Counter;
the high-order byte of the Program Counter IS set to zero. For example. the instruction
RST
IS

OOH

illustrated as follows:
Data
Memory

I I I

S.D.EC

-+

r ::::_ 2)
.>..

--I

~~-----_Il"'c
~~......
.......,

H.l
SP

5555

__ - / : .

~
+y-

pcl------~m".:m:;;;m~m~-----r~mmm

'r

:
RST

ssss-2

t-----I ssSS

Program
Memory

l.( :: ~r.....
---'......-...:C;.;.7--t : : : : + 1
t - - - - I mmmm + 2
t - - - - I mmmm + 3

OOH

~
76543210

4T

mm+1

It-_m_m_.... 5555 - 1

Restart instruction
Address code

3-13

Register Indirect
In standard indirect addressing. a memory location contains the effective address. and
the instruction specifies the address of the memory location containing the effective
address. In register indirect addressing. a register contains the effective address. and
the instruction specifies which of the registers contains the effective address. Note that
for a Load. for Instance. this IS Just another way of describing Implied addressing.
However. the ZSO has Jump instructions that allow a jump to the memory location
whose address is contained in the specified register. This is a form of indirect addressing. and is described separately because. while most microcomputers have Implied addressing. very few have register Indirect jumps.
The instruction
JP

(HL)

directs that a jump IS to be taken to the memory location whose address IS contained In
HL. This may be illustrated as follows:
S Z ACP/0 N C
F

Data

c:r::::o::I:D

B.C
D.E
H.l
SP
PC

pp

qq

mmmm

IX

IY

I
R

I
I

~
Program
mory

~
E9

JP IHLl

"":;('"'"
r.";-S-l4L.-II
;;-S- 4 "'3""--1112-.,......
~JUmpVlaHL

3-14

mmmm
mmmm+l
mmmm+2
mmmm+3

Immediate
Some texts Identify Immediate instructions as Memory Reference instructions. An Immediate instruction IS a 2-. 3-, or 4-byte instruction In which the last one or two bytes
hold fixed data that is loaded Into a register or memory location. The Z80 provides Immediate instructions to:

load 8-bit data into any of the 8-bit registers,

load 16-bit data into any of the register pairs or 16-bit registers,
store 8-bit data into any memory location using implied or indexed addressing,
perform arithmetic and logical operations using the Accumulator and 8-bit immediate data.
The Instruction
LD BC.OBCH

loads the immediate data value BC16 Into Register Pair BC. This may be illustrated as
follows:
S

F'

ZAcP/ON C

Data

I I I I I

B.C
D.E

, -

H.L

SP

Program
Memorv

::;;0::-::1 mmmm + 3

mmmm

PC

IX
IY

01

I
I

......

BC

00

mmmm
mmmm+l
mmmm+2
mmmm+3

LD

BC.OBCH

76543210

. ,. ,.,

4I~.~_"
'--------Regls'er Pair BC

76543210
1

Immediate data - low-order byte

Immediate data - high-order byte

3-15

Table 3-1. Frequently Used InstructIOns of the

zao

Instruction Code

Meaning

Add with Carry to Accumulator

Add
Logical AND
Call Subroutine
Call Conditional
Compare
Decrement
Decrement and Jump If Not Zero
Input
Increment
Jump Relative
Jump Relative Conditional
Load Reg ister
Load Accumulator Direct
Load Immediate
Store Register
Store Accumulator Direct
Move Register-to-Register
Output
Pop from Stack
Push to Stack
Return from Subroutine
Return Conditional
Rotate Accumulator Left Through Carry
Rotate Accumulator Right Through Carry
Shift Left Arithmetic
Shift Right Logical
Subtract

ADC
ADD
AND
CALL
CALL
CP
DEC
DJNZ
IN
INC
JR
JR
LD
LD
LD
LD
LD
LD
OUT
POP
PUSH
RET
RET
RLA
RRA
SLA
SRL
SUB

addr
cond,addr

cond,addr
reg,(HL)
A,(addr)
data
(HL),reg
(addrl.A
dst.src

cond

316

Table 3-2. Occasionally Used Instructions of the

Instruction Code
BIT
CPD.CPDR
CPI. CPIR
CPL
DAA
01
EI
EX
HALT
IND,INDR
INI. INIR
JP
JP
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LDD, LDDR
LDI. LDIR
NEG
NOP
OR
OUTD.OTDR
OUTI. OTIR
RES
RETI
RL
RLC
RLCA
RR
RRC
RRCA
SET
SRA
XOR

addr
cond,addr
A, (BC) or (DE)
HL,(addr)
reg. (xy+d isp)
rp,(addr)
xy.(addr)
(BC) or (DEl.A
(addr).HL
(xy+displ.reg
(addrl.rp
(addrl.xy
(HU,data
(xy+displ.data

zao

Meaning
Test Bit
Compare. Decrement. (Repeat)
Compare, Increment. (Repeat)
Complement Accumulator
Decimal Adjust Accumulator
Disable Interrupts
Enable Interrupts
Exchange
Halt
Input. Decrement. (Repeat)
Input. Increment. (Repeat)
Jump
Jump Conditional
Load Accumulator Secondary
Load HL Direct
Load Register Indexed
Load Register Pair Direct
Load Index Register Direct
Store Accumulator Secondary
Store HL Direct
Store Register Indexed
Store Register Pair Direct
Store Index Register Direct
Store Immediate to Memory
Store Immediate to Memory Indexed
Load, Decrement. (Repeat)
Load, Increment. (Repeat)
Negate (Twos Complement) Accumulator
No Operation
Logical OR
Output. Decrement. (Repeat)
Output. Increment. (Repeat)
Reset Bit
Retu rn from Interru pt
Rotate Left Through Carry
Rotate Left Circular
Rotate Accumulator Left Circular
Rotate Right Through Carry
Rotate Right Circular
Rotate Accumulator Right Circular
Set Bit
Shift Right Arithmetic
Logical Exclusive OR

3-17

Table 3-3. Seldom Used Instructions of the Z80

Meaning

Instruction Code
ADC
CCF
EXX
1M
RETN
RLD
RRD
RST
SBC
SCF
LD
LD
LD
LD
LD
LD

Add Register Pair with Carry to HL

Complement Carry Flag
Exchange Register Pairs and Alternatives
Set Interrupt Mode
Return from Non-Maskable Interrupt
Rotate Accumulator and Memory Left Decimal
Rotate Accumulator and Memory Right Decimal
Restart
Subtract with Carry (Borrow)
Set Carry Flag
Load Accumulator from Interrupt Vector Register
Load Accumulator from Refresh Register
Store Accumulator to Interrupt Vector Register
Store Accumulator to Refresh Register
Move HL to Stack Pointer
Move Index Register to Stack Pointer

HL.rp

A.I
A.R
I.A
R.A
SP.HL
SP.xy

ABBREVIATIONS
These are the abbreviations used in this chapter:
A.F.B.C.D.E.H.L

The 8-bit registers. A is the Accumulator and F is the Flag Word.

AF'.BC'.DE'.HL'

The alternate register pairs

addr

A 16-bit memory address

x(b)

Bit b of 8-blt register or memory location x

cond

Condition for program branching. Conditions are:

NZ - Non-Zero (Z = 0)
Z - Zero (Z = 1)
NC - Non-carry (C = 0)
C - Carry (C = 1)
PO - Parity Odd (p = 0)
PE - Panty Even (p = 1)
P - Positive Sign (S = 0)
M - NegatIVe Sign (S = 1)

data

An 8-blt bmary data unit

data16

A 16-bit binary data unit

disp

An 8-blt signed binary address displacement

xx(HU

The high-order 8 bits of a 16-bit quantity xx

Interrupt Vector register (8 bits)

IX IY

The Index registers (16 bits each)

label

A 16-blt Instruction memory address

xx(LO)

The low-order 8 bits of a 16-blt quantity xx

LSB

Least Significant Bit (Bit 0)

MSB

Most Significant Bit (Bit 7)

PC

Program Cou nter

port

An 8-blt I/O port address

3-18

pr

Any of the following register pairs:

BC
DE
HL
AF

The Refresh register (B bits)

reg

Any of the following registers:

A
B

C
D

E
H
L

rp

Any of the following register pairs:

BC
DE
HL
SP

SP

Stack POinter (16 bits)

xy

Either one of the Index registers (IX or IY)

Object Code

bbb

Bit number 000 (LSB) to 111 (MSB)

ccc

Condition Code

ddd

Destination register -

000 = non-zero
001 = zero
010 = no carry
all = carry
100 = panty odd
101 = parity even
110 = positive sign
111 = negative sign
same coding as rrr

ppqq A 16-bit memory address

rrr

Register

111
000
001
010
011
100
101

= A
= B
=C
= D
= E
=H
= L

sss

Source register -

same coding as rrr

Index register

xx

Register pair

0 = IX
1 = IY

xxx

00 = BC
01 = DE
10 = HL
11 = SP (rp) or AF (prl
Restart code (000 to 111)

yy

An 8-blt binary data unit

YYYY

A 16-blt binary data unit

3-19

Statuses

The
C
Z
S

zao has the following status flags:

PIO
AC
N

- Carry status
- Zero status
- Sign status
- Parity/Overflow status
- Auxiliary Carry status
- Subtract status

The following symbols are used In the status columns:

X
- flag IS affected by operation
(blank) - flag IS not affected by operation
1
- flag is set by operation
o
- flag IS reset by operation
U
- flag IS unknown after operation
P
- flag shows parity status
flag shows overflow status
I
- flag shows Interrupt enabled/disabled status

[[J]

Memory addressing: 1) the contents of the memory location

whose address IS contained In the designated register, 2) an
I/O port whose address IS contained In the designated register,

[J

The contents of a register or memory location,

For example:
[[HL]] -

[[HL]]

Indicates that the contents of the memory location addressed by

the contents of HL are incremented, whereas:
[HL] -

[HL]

Indicates that the contents of the HL register itself are incremented,

A

Logical AND

Logical OR

--

Logical ExclUSive-OR
Data IS transferred In the direction' of the arrow
Data IS exchanged between the two locations deSignated on either
side of the arrows,

3-20

INSTRUCTiON MNEMONICS

zao

Table 3-4 summarizes the

instruction set. The MNEMONIC column shows the
instruction mnemonic liN, OUT, LD). The OPERAND column shows the operands,
if any, used with the instruction mnemonic.
The fixed part of an assembly language instruction is shown in UPPER CASE. The
variable part !immediate data, I/O device number, register name, label or address)
is shown in lower case.
For closely related operands. each type is listed separately without repeating the
mnemonic. For instance. examples of the format entry

are:

LD

rp.(addrl
xy,(addr)

LD
LD

BC,(DAT21
IX,(MEM)

INSTRUCTION OBJECT CODES

The object code al1d instruction length in bytes are shown in Table 3-4 for each
instruction variation. Table 3-5 lists the object codes in numerical order.
For instruction bytes without variations, object codes are represented as two
hexadecimal digits (e.g., 3F).
For instruction bytes with variations in one of the two digits, the object code is
shown as one 4-bit binary digit and one hexadecimal digit (e.g., 11 x 1 D) in Table
3-5. For other instruction bytes with variations, the object code is shown as eight
binary digits (e.g., 01 sss001).

INSTRUCTION EXECUTION TIMES

Table 3-4 lists the instruction execution times in clock periods. Real time can be
obtained by dividing the given number of clock penods by the clock frequency. For
example, for an instruction that requires 7 clock penods, a 4 MHz clock will result In a
1.75 microsecond execution time.
When two possible execution times are shown !i.e., 5/11), it indicates that the
number of clock periods depends on condition flags. The first time is for "condition not met," whereas the second is for "condition met,"

STATUS
The

SIX

status flags are stored in the Flag register (FI as follows:

' - - t -....- i f - + - t - - -

These bits are not used

Carry status (carry out of bit 7)

' - - - - - - Subtract status

(1 after subtract operation, 0 otherwise)
' - - - - -_ _ Panty/Overflow
(for logical operations. 1 for even, 0 for odd panty.
For arithmetic. 1 for overflow)
' - - - - - - - - - - Auxiliary Carry status (carry out of bit 3)

' - - - - - - - - - - - - - Zero status (l for zero, 0 for nonzero)

' - - - - - - - - - - - - - - Sign status (value of bit 7)

3-21

In the individual instruction descriptions. the effect of instruction execution on

status is illustrated as follows:
S

AC Pia

1--+_+1-

1-

Modified to reflect results of execution

.....- - - - - Unconditionally reset to 0

Unconditionally set to 1

Unchanged
Unknown

L....

An X Identifies a status that IS set or reset. A 0 Identifies a status

that IS always cleared. A 1 Identifies a status that IS always set. A
blank means the status does not change. A question mark (7)
means the status IS not known.

3-22

STATUS

CHANGES
WITH
INSTRUCTION
EXECUTION

 Address Bus: AD-A7: [Cl

AB-A15: [B]

Type

Mnemonic

IN

Taole 3-4 A Summary of the

Operand

A,(portl

Object Code

DB yy

Bytes

Clock
Cycle.

zao Instruction Set

Status
Operation Performed:

PIO

AC

IA)-Iport)

10

Input to Accumulator from directly addressed I/O port

IN

INIR

reg.(CI

EO 01dddOOO

ED B2

11

20/15"

Address Bus: AO-A 7: port

AB-A15: I A)
Ireg)-[[CII
Input to register from 1/0 port addressed by the contents of C ..
If second byte is 70
only the flags will be affected
16
Repeat until I B) = 0:
[[ HLII
[[CII
IB]-IB] -1
I HL) - I HL) + 1
Transfer a block of data from I/O port addressed by contents of C
to memory location addressed by contents of HL. going from low

to)

addresses to high Contents of B serve as a count of bytes remaining tobe transferred.

N
to)

INOR

ED BA

20/15"

Repeat until [ B) = 0:
[[HLII-[[CII
IB)-IB] -1
[ HLJ - I HL) - 1
Transfer a block of data from 110 port addressed by contents of C
to memory location addressed by contents of HL, going from high
addresses to low Contents of B serve-as a count of bytes remaining
to be transferred

INI

ED A2

15

[[ HLI - I[ CII
IB)-IB].l
I HLJ -

[ HLI

+ 1

Transfer a byte of data from I/O port addressed by contents of C to

memory location addressed by contents of HL Decrement byte
count and increment destination address

 Address Bus: AO-A7: [C]

AS-AIS: [B]

Type

Mnemonic

Table 3-4. A Summary of the Z80 Instruction Set (Continued)

Operand

INO

Object Code

ED AA

Bytes

Clock
Cycles
15

Status
Operation Performed

PIO AC
?

N
1

[[HLl!-[[Cl!
IBI-IBI -1
[ HLI - I HLI - 1
Transfer a byte of data from 1/0 port addressed by contents of C to
memory location addressed by contents of HL Decrement both
byte count and destination address ,.

OUT

(portl,A

03 yy

11

Iport]-[Al
Output from Accumulator to directly addressed 110 port

OUT

(C),reg

ED 01 sssOOl

12

20/15

Address Bus: AO-A 7: port

AS-A1S: IAI
[[CIl-lreg]
Output from register to I/O port addressed by the contents of C"

OliR

(.0)

ED B3

'il
<I>

"Sc"

Repeat until I Bl = 0:
[[Cl!-[[HLl!
IBI-IBI -1
I HL] - [HLI + 1
Transfer a block of data from memory location addressed by con~

tents of HL to 110 port addressed by contents of C, going from low

memory to high Contents of B serve as a count of bytes remaining

to be transferred .

OTDR

ED BB

20/15

Repeat until I BI = 0:
[[Cl!-[[HLl!
IBI-IBI-l
[ HLI - [HLI - 1
Transfer a block of data from memory location addressed by con~
tents of HL to 110 port addressed by contents of C. going from high
memory to low, Contents of B serve as a count of bytes remaining
to be transferred, **

Address Bus: AO-A7: [CJ

A8-A15: [BJ
Type

Mnemonic

Table 3-4 A Summary of the

Operand

Object Code

Bytes

zao

Status

Clocl(
Cycles

Instruction Set (Continued)

Operation Performed

Pia AC

OUTI

ED A3

15

[[ C))
[[ HLI)
IB) -[B) - 1
[ HL)
I HLI + 1
Transfer a byte of data from memory location addressed by eontents of HL to JlO port addressed by contents of C Decrement byte

aUTO

ED AB

15

, , ,

[[ CIl - [[ HLI)
[ B) 1
I B)
[ HL) - [ HLI - 1
Transfer a byte of data from memory location addressed by contents of HL to 110 port addressed by contents of C Decrement both
byte count and source address

3A ppqq

13

'ij

-g"
~

c
0

LD

eN
I

A.laddrl

count and increment source address

[A)

[ addrl

Load Accumulator from directly addressed memory location

N
<.71

LD

HLladdri

2A ppqq

16

LD

rp.[addr)
xy.laddrl

ED 01xxl011 ppqq
l1xll101 2A ppqq

4
4

20
20

"uc

Load register pair or Index register from directly addressed memo-

"

a:

LD

[addrl.A

32 ppqq

13

LD

laddrl.HL

22 ppqq

16

LD

laddrl.rp
laddrl.xy

ED 01xx0011 ppqq
l1xl1101 22 ppqq

4
4

20
20

LD

A,IBCI
A.IDEI

OA
lA

1
1

:;"

>-

I;;

[ H) - [ addr + 11. [LI - [addr)

Load HL from directly addressed memory
[ rplHl1l
[addr + 1J. [rpILOI] - [ addrl or
[xyIHII] - [addr + 1J. [xylLOI) - [ addrl

.5

a:

ry
[addrl- [AI
Store Accumulator contents in directly addressed memory location
[ addr + 11 - [ HI. I addrJ - [ L)
Store contents of HL to directly addressed memory location
[ addr + 1] - [rpIHIIJ. [addrJ - I rplLO)] or
I addr + 1) - [ xylHIIJ. I addr] - [xyILO)]
Store contents of register pair or Index register to directly addressed memory
[A] - J[ BC]] or I AI - J[ DE]]
Load Accumulator from memory locati9n addressed by the contents of the specified register pair

Table 3-4 A Summary of the

Type

""c
E
.:!
,,_
"'..,
~ .~
E
" c
::!;g

Mnemonic

Operand

Object Code

Bytes

Clock
Cycles

zao Instruction Set (Continued)

Status
Operation Performed

PIO AC

LD

reg,(HLI

01ddd110

I reg] - IIHL]]

LD

(BCIA
(DEI. A

02
12

1
1

7
7

II BC]] - I A] or II DEll - I A]

(HLI.reg

01110sss

II HL]] - I reg]

Load register from memory location addressed by contents of Hl

LD

Store Accumulator to memory location addressed by the contents

of the specified register pair
Store register contents to memory location addressed by the con~

..5<:-

Ii:

LD
LD

reg.(xy+displ 11x11101 01ddd110

dlsp
(xy+disp),re9 11x1110101110sss
disp

19

19

tents
I reg] Load
II xy] +

of HL
II xy] + disp]
register from memory location using base relative addressing
dlsp] - I reg]

Store register to memory location addressed relative to contents of

Index register

CAl

LDIR

ED BO

20/16""

Ol

J:

":E

the contents of Hl to the memory location addressed by the con~

~
....

"a

iii

= 0:

Transfer a block of data from the memory location addressed by

'"..,"c
~c

Repeat until I BC]

((DE]] -((HLII
I DE] - I DE] + 1
I HL] - I HL] + 1
I BC] - I BC] - 1

LDDR

ED B8

20/16""

tents of DE, going from low addresses to high Contents of Be

serve as a count of bytes to be transferred
Repeat until I 8C] 0:
((DEli -((HLlI
I DE] - I DE] - 1
I HL] - I HLI ~ 1
18C] - I BC] - 1
Transfer a block of data from the memory location addressed by
the contents of HL to the memory location addressed by the con~
tents of DE. going from high addresses to low Contents of Be
serve as a count of bytes to be transferred

Table 3-4. A Summary of the

Type

Mnemonic

LDI

Operand

Object Code

ED AO

Bytes

Clock
Cycles

zao Instruction Set (Continued)

Status
Operation Performed

Pia AC
X

16

N
0

I[ DEll -If HLlJ

[DEI-[DE] + 1

I HLI - I HLI + 1
[ BC] - [ Bcl - 1
Transfer onB byte of data from the memory location addressed by
the contents of HL to the memory location addressed by the con-

tents of DE Increment source and destination addresses and decrement byte count

..

LDD

ED A8

16

CPIR

ED Bl

20/16"

'i:i
:l

~c
0

Co)

-.,j

'"..~

.
".

Ul

.!
"c
~

I[ DEll -I[ HLIl

[ DE] - [ DEI - 1
[ HL] - I HL] - 1
[ BC] - [ BC] - 1
Transfer onB byte of data from the memory location addressed by
ttie contents of HL to the memory location addressed by the 000tents of DE, Decrement source and destination addresses and byte
count
Repeat until [ A] = II HLlJ or [ BC] = 0:
I A] - I[ HLIl (only flags are affected)
[ HLI - [ HLl + 1
( BCI - [ BC] - 1
Compare contents of Accumulator with those of memory block ad-

"""
iii

dressed by contents of HL, going from low addresses to high.. Stop

when a match is found or when the byte count becomes zero

CPDR

ED 69

20/16"

Repeat until [A] = [[ HLJI or [ BCI = 0:

[ AI [[ HLJI (only flags are affected)
[ HLI - [ HLI - 1
[ Bel - [ BCI - 1
Compare contents of Accumulator with those of memory block ad~

dressed by contents of HL, going from high addresses to low Stop

when a match is found or when the byte count becomes zero

Table 3-4 A Summary of the

Type

Mnemonic

Operand

CPI

Object Code

ED A1

Bytes

Clock
Cycles

zao Instruction Set (Continued)

Status
Operation Performed

16

P/O

AC

"g"i5

= 1:

Compare contents of Accumulator with those of memory location

.!"o ':;::;CC
C

addressed by contents of HL Increment address and decrement

byte count

E.2
... J:

::<. u

ED A9

CPO

16

[AI - [I HLIJ (only flags are affectedI

[ HLI - [ HLl - 1
[ BCI - [ BCl - 1

u "
-o =
=
lQ

[AI - [I HLJI (only flags are affected)

[ HLl - [HLl + 1
[BCI - [ BCl - 1

(J)

Compare contents of Accumulator with those of memory location

addressed by contents of HL.. Decrement address and byte count

ADD
Co)
I

A,(HLl
A.(xy +disp)

86
11x11101 86 disp

19

[AJ -

dressing
[AJ - [AI + [I HLIJ + C or [ AJ - [Al + [I xyl + dispJ + C

iog
[AI - [AJ - [I HLIJ or [AI - [ AI - [I xyl + displ

co

ADC

A,(HU

A. (xy +displ

=
C

.
.!.
a::

SUB

>
l;

S8C

8E
11x111018Edisp

19

Add with Carry using implied addressing or base relative address-

.
.

::;
1::

AND

"u
C
0

(J)

OR

(HLI
(xy + displ

96
11x1110196disp

A.(HU
A. (xy + disp)

9E
11x111019Edisp

(HLI
(xy + disp)

A6
11x11101 A6disp

19

(fill

B6
11x11101 B6 disp

19

(xy + disp)

19

19

[AJ + [I HLIJ or [AI - [AI + [I xyJ + displ

Add to Accumulator using implied addressing or base relative ad-

Subtract from Accumulator using implied addressing or base rela-

tive addressing
[Al - [AI - [[ HLJI - C or [AJ -

[AI - [[ xyl + displ - C

Subtract with Carry using implied addressing or base relative addressing

[AI - [Al A [[ HLII or [AJ -

[AI A [[ xyl + dispJ

AND with Accumulator using implied addressing or base relative

addressing

[AI -

[AI V [[ HLII or [AJ -

[ AI V [[ xyl + displ

OR with Accumulator using implied addressing or base relative addressing

Table 3-4. A Summary of the Z80 Instruction Set (Continued)

Type

Mnemonic

Oparand

(HU

o .:
"
E

CP

,; g
8;
.

INC

>g

Ul8!.

W
I
N

Clock
Cycles

Status
Operation Performed
C

AE
l1xlll0l AE disp

1
3

7
19

(HU
(XV + displ

BE
llxlll0l BE disp

1
3

7
19

IHU
Ixv + disp)
(HU
(XV + displ

34
l1xl1101 34 disp
35
l1xlll0l 35 disp

1
3
1
3

11
23
11
23

(HU

CB 06

15

Ixv + disp)

l1xlll0l CB disp
06

23

PIO AC
P

N
0

[A] - [A].lJ.[[ HLJJ or [A] - [A].lJ.[[ xy]

+ disp]

Exc'usive~OR with Accumulator using implied addressing or base

relative addressing

:E g

Bytes

Ixv + disp)

XOR

~'i

Object Code

DEC

RLC

CD

[AI - [[ HLJJ or [A] - [[ xvi

+ disp]

Compare with Accumulator using implied addressing or base rela-

tive addrassin9. Onlv tha flags ara affected

[[ HLJJ - [[ HLJJ ... 1 or [[ XV] + disp] - [[ XV] + disp]

+1

Increment using implied addressing or base relative addressing

[[ HLJJ - [[ HLJJ - 1 or [[

xvi + disp] - [[ xvI + disp] - 1

Decrement using implied addressing or base relative addressing

&y7~OtJ
[[ HLJJ or [[

xvI + disp]

Rotate contents of memory location (implied or base relative address-

ing) laft with branch Carry

a:

"c
::'"

CB 16

15

l1xlll0l CBdisp
16

23

(HU

CB OE

15

Ixv + disp)

l1xl1101 CBdisp
OE

23

IHU

RL
(XV

+ disp)

:c

~7 .-oiJ
[[ HLJJ or [[ xvI

.
0

:E

RRC

+ disp]

Rotate contents of memory location left through Carry

Ul

l{7

~0J=LEJ

[[ HLJJ or [[ XV]

+ displ

Rotate contents of memory location right with branch Carry

Table 3-4. A Summary of the Z80 Instruction Set (Continued)

Type

Mnemonic

RR

Operand

IHLI
Ixy + disp)

Object Code

CB IE
llxlll0l CBdisp

Bytes

2
4

Clock
Cycles

15
23

Status
Operation Performed

Pia

AC

II HLJJ or [[ xy] + disp]

IE

'ii

SLA

"
'S"

IHLI
Ixy + disp)

CB 26
llxlll0l CBdisp

Rotate contents of memory location right through Carry

2
4

15
23

2
4

15
23

g
W
I

!
0

a:
c

SRA

."

IHLI
Ixy + disp)

::'"

CB 2E
llxlll0l CB disp

EJ.--17

<OIIliI

r--

[[ HLJJ or [[ xy] + disp]

Shift contents of memory location left and clear lSB (Arithmetic
Shift)

26

~7--'O~

7 --'0

[[ HLJJ or [[ XV] + disp]

2E

:E
f/l

>

j Shift

' (Arithmetic ShiftJ

:;

contents of memory location right and preserve MSB

::<'"

SRL

IHLI
(xy + disp)

CB 3E
llxlll0l CB disp
3E

2
4

15
23

0-..j7---"'0r---E]
[[ HLJJ or [[ xy] + disp]
Shift contents of memory location right and clear MSB (logical Shift)

Table 3-4. A Summary of the Z80 Instruction Set (Continued)

Type

Mnemonic

Operand

LD

reg, data

LD

E
E

LD

rpdata16
xV.data16
(HLI.data
(xv'+ disp).
data

..,.!!.

Object Code

00ddd110

vv

Bytes

Clock
Cycles

3
4
2
4

10
14
10
19

Status
Operation Performed
C

P/O

AC

N
[reg] -data
Load immediate into register

OOxxOOO 1 vvvv
11x11101 21 vvvv
36 VV
l1xl110136disPVY

[rpl - data16 or 1xyl - data16

Load 16 bits of immediate data into register pair or Index register

[[ HLiI - data or [[ xvi

+ dispi - data

Load immediate into memory location using implied or base relative

addressing

JP

label

C3 ppqq

10

[PCI -label

JR

disp

18 (disp-2)

12

[PCI - I PCI

Jump to instruction at ~ddress represented by label

"-

..,:>

JP

c..>
I
c..>
CALL

(HLI
(XV)

E9
11x11101 E9

1
2

4
8

label

CD ppqq

17

..
..,
.
1i

cond.label

11ccc100 ppqq

10117

RET

C9

10

11 cccOOO

5/11

'iii

.,

(,)

:S:>

e
:>

(/J

Jump to address contained in HL or Index register

[[ SP] - 1I - 1PC[HII]
[[ SPI - 21 - [ PC(LO))

Jump to subroutine starting at address represented by label

CALL

.c

[ PCI - 1HLI or 1PCI - [ xvi

1SPI - 1SPI - 2
[PCI -label

a::

+ 2 + (disp-Z)

Jump relative to present contents of Program Counter

Jump to subroutine if condition is satisfied; otherwise. continue in

sequence
1PC(LO)) - [[ SPJI
[PC(HIII - [[ SPI + 11
1SPI - 1SPI + 2
Return from subroutine

RET

cond

Return from subroutine if condition is satisfied; otherwise. continue

in sequence

Table 3-4 A Summary of the Z80 Instruction Set (Continued)

Type

Mnemonic
ADD

Operand

Adata

Object Code

C6

yv

Bytes

Status

Clock
Cycles

PIO

AC

Operation Performed

I Al - 1Al

+ data

Add immediate to Accumulator

ADC

A.data

CE

vv

1Al - 1Al

+ data + C

Add immediate with Carry

.
...
.

Co

SUB

data

06

vv

1Al - 1Al - data

SBC

A.data

DE

vv

[ Al -

Subtract immediate from Accumulator

'6
E
E

w
(.,

[ Al - data - C

Subtract immediate with Carry

AND

data

E6

vv

1Al - I Al to. data

OR

data

F6

vv

[ Al -

AND immediate with Accumulator

[ Al V data

OR immediate with Accumulator

XOR

data

EE

vv

[AI- [AI-\l-data

CP

data

FE

vv

IAl - data
Compare immediate data with Accumulator contents; only the
flags are affected

JP

cond.label

10

Exclusive-OR immediate with Accumulator

11 cccOl 0 ppqq

If condo then 1PCl - label

Jump to instruction at address represented by label if the condition

is true

JR

C.disp

38 (disp-21

7/12

'6
c

JR

NC.disp

30 (disp-2)

7/12

::l

1PCI - I PCI

+ 2 + (disp - 2)

If C

= O. then

[ PCl - [PCl

+ 2 + (disp -21

Jump relative to contents of Program Counter if Carry flag is reset

JR

Z,disp

28 (disp-2)

7/12

If Z

7/12

If Z

0
Co

..,E

= 1. then

Jump relative to contents of Program Counter if Carry flag is set

()

If C

= 1. then

1 PCl - [PCl

+ 2 + (disp -21

Jump relative to contents of Program Counter if Zero flag is set

JR

NZ.disp

20 (disp-21

= O. Ihen

[ PCl - I PC]

+ 2 + (disp -21

Jump relative to contents of Program Counter if Zero flag is reset

DJNZ

disp

10 (disp-2)

8/13

181-181-1
If [81 ,pO. then 1PCl

+2 +

(disp ~'2)

Decrement contents of B and Jump relative to contents of Program

Counter if result is not O.

Table 3-4. A Summary of the l80 Instruction Set (Continued)

Type

Mnemonic

LD

Operand

dst.src

Object Code

01dddsss

Bytes

Clock
Cycles

Status

Operation Performed

PIa

Ac

I dstl - [ src!
Move contents of source register to destination register Register
designations src and dst may each be A, B. C, D, E, H or L

LD

A,I

ED 57

IA!-Ill

LD

A,R

ED 5F

[A!-[R)

LD

I,A

ED 47

[II-[A!

LD

RA

ED 4F

[R!-[A!

LD

SP,HL

F9

[SP!-[HLI

::;;

LD

SP,xy

11x11101 F9

10

[SP) - [ xy)

EX

DE,HL

EB

[DE) - - [HL)

'Z'"

EX

AF,AF'

OB

IAFI - - IAF')

09

Move contents of Interrupt Vector register to Accumulator

Move contents of Refresh register to Accumulator
Load Interrupt Vector register from Accumulator
Load Refresh register from Accumulator

w,

w
w

'0"

Move contents of HL to Stack Pointer

>

"til

'c,

a:'"

Move contents of Index register to Stack Pointer

Exchange contents of DE and HL

Exchange program status and alternate program status

EXX

CBC)
[DE!
[HL)

- - CBC')
IDE'!
[HL')

Exchange register pairs and alternate register pairs

Table 3-4 A Summary of the Z80 Instruction Set (Continued)

Type

Mnemonic

ADD

Operand

A,reg

Object Code

10000",

Bytes

Clock
Cycles

Status
Operation Performed

P/O AC
0

I AI - [AI + [regl
Add contents of register to Accumulator
[ AI - [AI + [regl + C
[ AI -

ADC

A,reg

10001",

SUB

reg

10010",

[ AI - [AI - [regl - C
I AI - I AI A [regl
[ AI -

Add contents of register and Carry to Accumulator

[ AI - I regl

Subtract contents of register from Accumulator

SBC

A,reg

10011",

Subtract contents of register and Carry from Accumulator

AND

reg

10000",

OR

reg

10110",

AND contents of register with contents of Accumulator

[ AI V [ regl

OR contents of register with contents of Accumulator

W
.j>.

"a.

XOR

reg

10101rrr

[Aj-[AI>J.[ regl

IAI -Iregl

Exclusive-OR contents of register with contents of Accumulator

CP

reg

10111",

Compare contents of register with contents of Accumulator Only

the flags are affected

[ Hli - [fill + [rpl

[ HLI - [ Hli

';'

"

ADD

Hl,rp

00 1001

11

'63

ADC

Hl,rp

ED 01 1010

15

16-bit add register pair contents to contents of HL

"

+ [rpl + C

16-bit add with Carry register pair contents to contents of Hl

0:

SBC

Hl,rp

ED 01 0010

15

[ Hli - [ HLI - [rpl - C

Hl
[ IXI - [ IXI

l6-bit subtract with Carry register pair contents from contents of

ADD

IX,pp

DO 00 1001

15

+ [ppl

l6-bit add register pair contents to contents of Index register IX

ADD

IY,rr

FD 00 1001

15

Ipp = BC, DE, IX, SP)

[IYI-[IYI + [rrl
l6-bit add register pair contents to contents of Index register IV

Irr = BC, DE, IY, SPI

Table 3-4. A Summary of the

Type

Mnemonic

Operand

Object Code

Bytes

Clock
Cycles

DAA

27

CPL

2F

..

NEG

ED 44

INC

reg

00rrrl00

~
.s,
a::

INC

rp
XV
reg

00xx0011
l1xl1101 23
00",101

1
2
1

6
10
4

rp
xv

00xxl0ll
l1xl110128

1
2

6
10

07

zao Instruction Set (Continued)

Status
Operation Performed

P/O

AC

N
Decimal adjust Accumulator, assuming that Accumulator contents are
the sum or difference of BCD operands

[AI-[AI
Complement Accumulator (ones complement)

[AI-IAI + 1

I reg] - [regl + 1

Negate Accumulator (twos complement)

c-

..

Increment register contents

DEC

I rpl- [rpl + 1 or [XV] - [xV] + 1

Increment contents of register or Index register

[ reg] - [ reg] - 1
Decrement register contents

DEC

W
I

[rp] - [rpl - 1 or [XV] - [xvi - 1

Decrement contents of register pair or Index register

<.n
RLCA

i
c

RLA

17

.:::

:<:
rJ)

..

[A]

[AI

Rotate Accumulator left through Carry

's,
a::

lIIII

Rotate Accumulator left with branch Carry

..,a::

..

of:l
~7 .-of:l
~7

RRCA

OF

li7---'0lJ.8
[A]
Rotate Accumulator right with branch Carry

Table 3-4 A Summary of the Z80 Instruction Set (Continued)

Typel Mnemonic

RRA

Operand

Object Code

Byte.

Clock
Cycle.

1F

Operation Performed

L:f7---'0~

I OlD I

(A!

Rotate Accumulator right through Carry

RLC

reg

CB OODODm

~7~OP

..

;;

( regl
Rotate contents of register left with branch Carry

"cc
;;

~7~DP

to)

W
0)

I a..

reg

Rl

CB 00010rrr

( reg!

IX:

'"..

Rotate contents of register left through Carry

::
:E
til

0>
IX:

I
I

reg

RRC

C80DODlm

D
( regl
Rotate contents of register right with branch Carry

RR

reg

CB 00011rrr

L:f7 - - - . 0
( regl
Rotate contents of register right through Carry

SLA

reg

CB 00100m

I I IIIIIII
2

&--i7~0~0
[ regl

Shift contents of register left and clear LSB (Arithmetic Shift)

Table 3-4. A Summary of the

Type

Mnemonic

SRA

Operand

reg

Object Code

CB 00101rrr

Bytes

Clock
Cycles

zao Instruction Set (Continued)

Status
Operation Performed

S 'P/O

AC

6~0~
I reg]

Shift contents of register right and preserve MSB (Arithmetic Shift)

SRL

reg

CB00111m

""c

.....

-...J

~a

~O~

I regl
Shift contents of register right and clear MSB (Logical Shift)

'i!i

CAl
I
CAl

RLD

ED 6F

18

11

;";

I3
[AI

II:

I
01

'tl
C

::"

I3 oI
I ~

11

Rotate one BCD digit left between the Accumulator and memory loca-

:c

tion (implied addressing) Contents of the upper half of the Accumulator are not affected

'"

i
a:

RRD

ED 61

18

11

I3
[AI

r
o

t Il
11 4 I 3 o 1
[[HLJI

Rotate one BCD digit right between the Accumulator and memory

location (implied addressing) Contents of the upper half of the Accumulator are not affected.

Table 3-4.. A Summary of the Z80 Instruction Set (Continued)

Type

Mnemonic

Operend

Object Code

Bytes

Clock
Cycles

Status
Operation Performed

P/O AC

BIT

b,reg

CB 01bbbrrr

Z -reglbl

BIT

b,lHLI
b,(xy -l displ

CB 01bbb110
11x11101 CB disp
01bbb110
CB 11 bbbrrr

2
4

12
20

Z-

reg(bl-1

CB 11bbb110
l1x11101 CB disp
11bbb110
CB 10bbbrrr

2
4

15
23

I[ HLIJ(bl -

reg(b) - 0

C810bbb110
11x11101 CBdisp
10bbb110

2
4

15
23

[[ HL]J(b) - 0 or [[ xvI + dispJ(bl - 0

11xx0101
11x11101 E5

1
2

Zero flag contains complement of the selected register bit

r:
0

"
C

..
:;;

0.

iii

SET

b,reg

SET

b,lHLI
b,!xV + displ

[[ HLJI(b) or Z - [[ xyl + dispJ(b)

Zero flag contains complement of selected bit of the memory laeation (implied addressing or base relative addressing)

Set indicated register bit

RES

b,reg

RES

b,lHLI
b. (XV + displ

1 or [[ xyl + dispJ(bl -

Set indicated bit of memory location (implied addressing or

base relative addressing,
Reset indicated register bit

CAl
I
CAl

co

PUSH

pr
xy

Reset indicated bit in memory location (implied addressing or base

relative addressing)

11
15

[[ SPI-11 - [pr(HI}]
[[ SPI-2] - [pr(LOI]
I SP] - [ SPI-2
Put contents of register pair or Index register on top of Stack and
decrement Stack Pointer

POP

'"

~"

pr
XV

11xxOO01
11x11101 E1

1
2

10
14

[ pr(LO}] - [[ SPJI
Ipr(HII]-[[SP! + 11
I SPI - [SPI + 2
Put contents of top of Stack in register pair or Index register and
increment Stack Pointer

EX

ISPI,HL
(SPl,xv

E3
11x11101 E3

1
2

19
23

[H]--[[SPI+1!
[LJ-- [[SPIJ
Exchange contents of HL or Index register and top of Stack

Table 3-4 A Summary of the Z80 Instruction Set (Continued)

Type

Mnemonic

01
EI
RST

Operand

Object Code

F3
FB
llxxxl11

Bytes

Clock
Cycles

1
1
1

4
4
11

2
2
2
2
2

14
14
8
8
8

Status
Operation Performed

PIa AC

N
Disable interrupts
Enable interrupts

1I SPI-1 I - [PCIHIII
1I SPI-21 - [ PCllO)]
[ SPI - [ SPI-2
[PCI -18.n)16

11

Restart at designated location

Return from interrupt
Return from nonmaskable interrupt
Set interrupt mode 0,. 1. or 2

Co)

w
co

RETI
RETN
1M

0
1
2

ED
ED
ED
ED
ED

40
45
46
56
5E

..

SCF

37

l!!

CCF

3F

Nap
HALT

00
76

1
1

4
4

a
f/)

Execution time shown is for one iteration

C-l
Set Carry flag
C-C
Complement Carry flag
No operation -

volatile memories are refreshed

CPU halts. executes NOPs to refresh volatile memories

Table 3-5. Instruction Object Codes in Numencal Order

OBJECT CODE

OBJECT CODE

INSTRUCTION

INSTRUCTION

39

ADD

HL,SP

00

Nap

01 yyyy

LD

BC,data16

3A ppqq

LD

A,laddrl

02
03

LD
INC

IBCl.A
BC

3B

DEC

SP

04

INC

3C
3D

INC
DEC

A
A

05
06 yy

DEC

3E yy

LD

Adeta

LD

B.data

3F

CCF

07

RLCA

40sss

LD

B.reg

08

EX

AF.AF

46

LD

B.lHL)

09

ADD

HL,BC

4 lsss

LD

C,rag

OA

LD

A,IBCl

4E

LD

C.IHLl

OB

DEC

BC

50sss

LD

D.reg

OC

INC

56

LD

D,IHU

00

DEC

5 lsss

LD

E,reg

OE yy

LD

C,data

5E

LD

E,IHL)

OF

RRCA

6 Osss

LD

H.reg

10 disp-2

DJNZ

disp

66

LD

H.IHU

11 yyyy

LD

DE.data16

6 lsss

LD

L.reg

12

LD

IDElA

6E

LD

L,IHLl

13

INC

DE

10sss

LD

IHLl.reg

14

INC

76

HALT

15

DEC

7 1555

LD

A,reg

16 yy

LD

D.data

7E

LD

A.lHU

17

RLA

80m

ADD

A.reg

18 disp-2

JR

disp

86

ADD

A,IHL)

19

ADD

ADC
ADC

lB

LD
DEC

HL.DE
A,IDE)

81m

lA

DE

90m

SUB

A,IHU
reg

lC

INC

96

SUB

IHL)

10
IE yy

DEC

91m

SBC

A,reg

LD

E,data

9E

SBC

IF

RRA

AOrrr

AND

A.IHU
reg

20 disp-2

JR

NZ,disp

A6

AND

21 yyyy

LD

HL,datalf

Alm

XOR

IHU
reg

22 ppqq

LD

laddrl.HL

AE

XOR

IHL)

23

INC

HL

BOrn

OR

reg

24

INC

B6

OR

25
26 yy

DEC

B 1m

CP

IHLI
reg

LD

H.data

BE

CP

IHU

27

DAA

CO

RET

NZ

28 disp-2

JR

Cl

POP

BC

29

ADD

Z.disp
HL,HL

C2 ppqq

JP

NZ.addr

2A ppqq

LD

HL,(addrl

C3 ppqq

JP

addr
NZ,addr

8E

A,reg

2B

DEC

HL

C4 ppqq

CALL

2C

INC

C5

PUSH

BC

20

DEC

C6 yy

ADD

A,data

L,data

C7

RST

OOH

C8

RET

2E

LD

2F

CPL

30 disp-2
31 yyyy
32 ppqq

JR

NC,disp

C9

RET

LD
LD

SP.data16
laddrl.A

CA ppqq
CBOOm

JP
RLC

reg
IHL)

Z,addr

33
34

INC

SP

CB 06

RLC

INC

IHLl

CBO 1m

RRC

reg

35

DEC

IHU

CB OE

RRC

IHL),data

36 yy

LD

37

SCF

38

JR

C.disp

3-40

CB 1 Om

RL

IHU
reg

CB 16

RL

IHL)

CB 11m

RR

reg

Table 3-5. Instruction Object Codes

OBJECT COOE
RR

CB 2 Orrr

SLA

CB 26
CB 2 lrrr

SLA
SRA

CB 2E
CB 3 lrrr

SRA

CB 01bbbrrr
CB 01bbbll0
CB 10bbbrrr
CB 10bbbll0
CB llbbbrrr
CB llbbbll0
CC ppqq
CD ppqq
CE yy

Numencal Order (Continued)

INSTRUCTION

CB IE

CB 3E

In

SRL
SRL
BIT
BIT

RES

DO CB disp l1bbbl10

IHLI
reg
IHL)

DO El
DO E3

SET
POP

DO E5

EX
PUSH

reg

DO E9

JP

IHLI

DO F9
DE yy

LD
SBC

OF
EO

RST
RET
POP
JP

b,reg

b.IHL)
b.IHLI

SET

b,reg

SET
CALL

b.IHLI

CF
DO
01

RET
POP

02 ppqq
03 yy

JP
OUT

04 ppqq
05
06 yy

CALL
PUSH

INSTRUCTION

DO CB disp 10bbbl1O

RES
RES

CALL
ADC
RST

OBJECT CODE

IHLI
reg

b,reg

El
E2 ppqq
E3
E4 ppqq
E5
E6 yy

Z.addr
addr
A.data
OBH
NC

E7
E8
E9

DE
NC.addr
IportlA
NC.addr
DE

EX
CALL
PUSH
AND
RST
RET

EA ppqq
EB

JP
JP
EX

EC ppqq
ED 01dddOOO

CALL
IN

ED 01sss001
ED 01xx 2

OUT
SBC
LO
NEG

b,(lX+displ
b,(lX+disp)
IX
ISPI.IX
IX
(IX)
SP,IX
A,data
18H
PO
HL
PO,addr
ISP),HL
PO,addr
HL
data
20H
PE
(HL)
PE.addr
DE,HL
PE,addr
reg.lCI
ICl.reg
HL,rp

data

07

SUB
RST

DB
09

RET
EXX

EO 01xx 3 ppqq
ED 44

DA ppqq
DB yy

JP
IN

C.addr
A.lpert)

ED 45

RETN

CALL

C.addr

ED 010nnl1O
ED 47

1M

DC ppqq

LO

I.A

DO OOxx 9
DO 21 yyyy
DO 22 ppqq

ADD

IX.pp
IX.data16
laddrl.IX

ED 01xx A
ED 01xx B ppqq
ED 40

AOC
LD
RETI

HL,rp
rp,laddri

IX
IX.laddri

ED 4F
ED 57

LD
LD

IX
(IX + displ
(IX + displ

ED 5F
ED 67

LD
RRD
RLO

RA
A,I
A,R

DO 23
DO 2A ppqq

LD
LD
INC
LD

DO 2B
DO 34 disp

DEC
INC

DO 35 disp

DEC
LD

DO 36 disp yy
DO 01dddll0 diso

10H

ED 6F
ED AO

(IX + displ.data
reg.(lX + disp)

DO -' 0555 disc

DO 86 disp
DO 8E disp

ADD
ADC

DO 96 disp
DO 9E disp

SUB
SBC

(IX + displ

ED A8
ED A9

A,lIX + disp)

ED AA

CPO
iNO

DO A6 disp
DO AE disp

AND
XOR

(lX+displ
(lX+displ

ED AB
ED 80

LOIR

DO B6 disp

OR

DO BE disp

CP

(IX + displ
(IX + displ

ED Bl
ED B2

CPIR
INIR

DO CB disp 06

RLC

OTIR

ED B8

LOOR

DO CB disp 16

RRC
RL

(IX + diSPI
(IX + displ

ED B3

DO CB disp OE

(lX+displ

ED B9

CPDR

DO CB disp IE

RR

(IX +dispi

ED BA

INOR

DO CB disp 26

SLA

(lX+dispi

DO CB disp 2E
DO CB disp 3E
DO CB disp 01bbbl10

SRA
SRL
BIT

(IX + displ
(lX+displ

ED BB
EE yy
EF

OTOR
XOR
RST

(IX + dispI,reg
A,(lX+ disp)
A,(lX+disp)

b,(lX+displ

3-41

LDI
CPI
INI

LD
LD

ED Al
ED A2
ED A3

laddrl.rp

oun
LOO

OUTO

data
28H

Table 3-5. Instruction Object Codes in Numerical Order (Continued)

OBJECT CODE

INSTRUCTION

OBJECT CODE

INSTRUCTION

FO

RET

FD 8E disp

ADC

A,liY+disp)

Fl

pop

AF

FD 96 disp

SUB

liY+disp)

F2 ppqq

JP

P.addr

FD 9E disp

SBC

A,lIY+ disp)

F3

01

FD A6 disp

AND

lIY+disp)
lIY+ displ

F4 ppqq

CALL

P.addr

FD AE disp

XOR

F5

PUSH

AF

FD B6 disp

OR

liY+displ

F6 yv

OR

data

FD BE disp

CP

lIY+ displ
liY+ displ

F7

RST

30H

FD CB disp 06

RLC

F8

RET

FD CB disp OE

RRC

lIY+ disp)

F9

LD

SP.HL

FD CB disp 16

RL

liY+dispJ

FA ppqq

JP

M.addr

FD CB disp lE

RR

IIY+disp)

Fa
Fe ppqq

EI

FD CB disp 26

SLA

lIY+ disp)

FD CB disp 2E

SRA

IIY+displ

CALL

M.addr

FD ooxx 9

ADD

IY.IT

FD CB disp 3E

SRL

lIY+dispi

FD 21 yvyv

LD

IY.data16

FD CB disp 01bbbll0

BIT

b,liY+ dispi

FD 22 ppqq

LD

laddrJ.IY

FD CB disp lObbbl10

RES

b,liY+ disp)

FD 23
FD 2A ppqq

INC

IY
IY,(addr)

SET
POP

IY

FD CB disp llbbbll0
FD El
FD E3

b,liY+disp)
IY
(SPI,IV

FD 2B

LD
DEC

FD 34 disp

INC

lIY+ displ

FD E5

PUSH

IY

FD 35 disp

DEC

lIY+ displ

FD E9

JP

lIY)

FD 36 disp yv

LD

lIY + dispi,data

FD F9

LD

SP.lY

FD 01ddd 110 disp

LD

rag.lIY + displ

FE yv

CP

data

FD 7 Osss disp

LD

lIY + dispi,reg

FF

RST

38H

FD 86 disp

ADD

A,lIY+disp)

3-42

EX

ADC A,data-ADD IMMEDIATE WITH CARRY TO

ACCUMULATOR
S ZACP/ON l
mxlxlxlolx'

Data

~c+xx+~

xx

B.C
D.E
H.L

SP

~-:i""'mmmm+D

mmmm

PC

IX

~
Program

Memory

IV
I
R

I
I

CE
yy

mmmm
mmmm+ 1

mmmm+2
mmmm+3

--

ADC A.

data

CE

vv

-.-

Add the contents of the next program memory byte and the Carry status to the Accumulator.
Suppose xx=3A16. VV=7C16. and Carry=O. After the Instruction
ADC

A.7CH

has executed. the Accumulator will contain B616:

3A
7C
Carry

~"Sto

0011
01 11

1010
1 100
0
0110

fJ

1011
1

No carry. set C to 0

0...,.1 =1. set PIO to 1

LNoo-"ro

""'t ~t Z to 0

Carry. set AC tb 1
Addition instructIOn. set N to 0

The ADC Instruction is frequentlv used in multibyte addition for the second and subsequent bytes.

3-43

ADC A,reg -

ADD REGISTER WITH CARRY TO

ACCUMULATOR

ZACP/ON!

r,

tillIillI2IEl

Data

xx+yy+C

"-

} ".10'

xx

,.

B.C
O.E

-..A,B,C,D,E,H
orLisyy

H.L

SP
PC

__---::I mmmm +

mmmm

IX
IY

I
R

~
Program

Memory

10001 xxx

mmmm

t-_ _-i\mmmm + 1

1-_ _--1mmmm + 2
1-_ _--1mmmm + 3
ADCA

reg

'-v-'

10001

xxx

000

for
for
for
for
for
for
for

001
010
011
100
101
111

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Add the contents of Register A, B, C. D. E. H or L and the Carry status to the Accumulator.
Suppose xx=E316. Register E contains A016' and Carry= 1. After the instruction
ADC

A.E

has executed. the Accumulator will contain 8416:

E3
AO
Carry

1 1 10
1010
1000

1 sets S to 1
Carry, set C to 1

001 1
0000
1
0100

fU
'

LN~-"" ~'"It '" z

10 0

No carry, set AC to 0

1"'1-1 =0, set P/O to 0

Addition instruction. set N to 0

The ADC instruction is most frequently used in multibyte addition for the second and
subsequent bytes.

3-44

ADC A,(HL) -ADD MEMORY AND CARRY TO

ADC A,(lX+disp)
ACCUMULATOR
ADC A, (ly+disp)
S ZACP!ON C
F IXIXIX

Data
Memorv

XIO XI

....l..

'::x.+yy+~

~x

yy

ppqq

B.C
D.E
H.L

qq

PP

SP
PC

mmmm

..l~mmm+V

IX
IY
I
R

Program
Memory

~
BE

The illustration shows execution of ADC A(HL):

ADC A.(HL)
---.-

mmm
mmm m+1
mmmm+2
mmmm+3

8E

Add the contents of memory location (specified by the contents of the HL register pair)
and the Carry status to the Accumulator.
Suppose xx=E316. yy=A016. and Carry=l. After the instruction
ADC

A.(HL)

has executed. the Accumulator will contain 8416:

E3
AO
Carry

1 1 10
1010

001 1
0000
1
0100

fJ

1000
, 00" S 10 ,

Carry. set C to 1

N,"_""

,,,"Ie '"

Z 10 0

No carry. set AC to 0

1 >f 1=0, set Pia to 0

Addition Instruction. set N to 0

ADC A.(IX+disp)
--..--.DD

8E

Add the contents of memory location (specified by the sum of the contents of the IX
register and the displacement digit dl and the Carry to the Accumulator.
ADC A(lY+disp)
--..--.-

FD

8E

This instruction is Identical to ADC A.(IX+displ. except that It uses the IY register instead of the IX register.
The ADC instruction is most frequently used In multibyte addition for the second and
su bseq uent bytes.
3-45

ADC Hl,rp S

ADD REGISTER PAIR WITH CARRY TO HAND l

ZACP/ON

Data

FWX (xlx(o Ix.

Be. DE. HL or SP

' " contain yyyy

B.C
D.E
H.L
SP
PC

xx

~xxx+yyyy

xx

......

mmmm

IX
IY
I

+c

Program

Memorv

~mmm+2

mmmm
mmmm + 1
mmmm+2
t - - -...... mmmm + 3
ED

01xxl0l0

l2t
ED 01

X X

1010

00
01
10
11

for
for
for
for

-.-

rp IS register pair Be
rp is register pair DE
rp is register pair HL
rp IS Stack Pointer

Add the 16-blt value from either the BC. DE. HL register pair or the Stack POinter. and
the Carry status. to the HL register pair.
Suppose HL contains A53616. BC contains 104416. and Carry=l. After execution of
ADC

HL.BC

the HL register pair will coptaln:

A536
1044
Carry

1 " " 510 1

1010010100110110
0001000001000100
1
1011010101111011

No carry. set C to 0 " ' (

Noo -rero re,,1t '" Zto a

No carry. set AC to 0

0"'"0=0. set P/O to 0

Addition instruction. set N to 0

The ADC Instruction is most frequently used In multibyte addition for the second and
subsequent bytes.

3-46

ADD A,data - ADD IMMEDIATE TO ACCUMULATOR

s

Z AcP/o N C

:mmrmm

=--:'"~>-

xx

B.C
D.E

H.L
SP

PC

.;&~mmm+!)

mmmm

IX

IY
I
R

I~~" I
Data

I
I

Program
Memory
C6
yy

--...-.

-.-

C6

yy

mmmm
mmmm+l
mmmm+2
mmmm+3

ADD A. data

Add the contents of the next program memory byte to the Accumulator.
Suppose xx=3A 16. yy= 7C 16. and Carry=O. After the instruction
ADD

A.7CH

has executed. the Accumulator will contain 8616:

3A
001 1
7C = 0 1 1 1

101 1
1 " " S to 1
No carry. set C to 0

1010
1100
01 10

Oolf 1=1. set P/O to 1

LNoo."rn re'"It. '" Z to 0

Carry. set AC to 1
Addition instructIOn. set N to 0

This is a routine data manipulation instruction.

3-47

ADD A,reg S

Ac P/O

ADD CONTENTS OF REGISTER TO

ACCUMULATOR
N C

Data

I x I x I x I x I 0 Ixl

f.

xx

yy
' " x::- )

} ~Io'

B.C

--'A.B.C.D,E.

D,E

HorL,syy

- +v

H,L

SP
PC

mmmm

IX
IV
I

- ---:;;'mmm

Program
Memory

10000xxx

mmmm
mmmm+1
1---""'""1 mmmm + 2

I
I

~
t - - - - - t mmmm + 3

ADD
-..--'

reg

-.-

10000 xxx

000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Add the contents of Register A, B. C. D, E, H or L to the Accumulator,

Suppose xx=E316, Register E contains A016' After execution of
ADD A.E
the Accumulator will contain 8316:
E3
AO =

1 " " S '0 1

Carry, set C to 1

1 1 10
10 10
1000

001 1
0000
0011

r
-

1 1=0. set P/O to 0

This

IS

LNoo."m '''""_ '"

Z 10

No carry. set AC to 0
Addition Instruction. set N to 0

a routine data manipulation instruction

3-48

ADD A,(HL)-ADD MEMORY TO ACCUMULATOR

ADD A, (IX+disp)
ADD A,(lY+disp)
S

Z AC Pia N C

Data

x X X X a X
F~
A

_/ r-

~......

xx

B.C

xx+yy

o.E
H.L
SP

PC

mmmm

IX

ppqq

L-."'""':I'mmmm +
~

......

Memory

YY

ppqq + d

Program
Memorv

IY
I
R

I
L.0pqq +

d)"=

00
86
d

mmm m
mmm m+ 1
mmm m+2
mmm m+3

The illustration shows execution of ADD A.(lX+disp).

ADD A.(lX+disp)

~-,.-

DO

86

Add the contents of memory location (specified by the sum of the contents of the IX
register and the displacement digit d) to the contents of the Accumulator.
Suppose ppqq=400016. xx= 1A16, and memory location 400F16 contains 5016. After
the instruction
ADD A.(lX+OFH)
has executed, the Accumulator will contain 6A16.
lA
0001
50 = 0 1 0 1
01 10

1010
0000
1010

fJ

0"" S ,,0
No carry. set C to 0

LNOO~";O "w". ,," ,,0

No carry, set AC to 0

0"9"0=0; set PIO to 0

Addition instructiOn. set N to 0

~+~
FD

86

This Instruction IS Identical to ADD A. (lX+displ. except that It uses the IY register In
stead of the IX register.
ADD A,(HU
~

86
This version of the instruction adds the contents of memory location. specified by the
contents of the HL register pair. to the Accumulator.
The ADD Instruction IS a routine data manipulation instruction.

3-49

ADD HL,rp -

ADD REGISTER PAIR TO HAND L

S Z ACP/D N C

Data

FI ! Ix I 10 Ix I
BC. DE. HL or SP
contain VYVY

A
B.C
D.E
H.L

xx

SP
PC

"....,

....---:l. xxxx + YVYV

xx

......

mmmm

IX
IV
I

I.

Program

Memory

~mmm+1

00xx1001 mmmm
mmmm

1)t
00 xx 1001
OOfor
01 for
10 for
11 for

rp IS register pair BC
rp is register pair DE
rp is register pair HL
rp IS Stack POinter

Add the 16-blt value from either the BC. DE. HL register pair or the Stack Pointer to the
HL register pair.
Suppose HL contains 034A16 and BC contains 214C16. After the Instruction
ADD HL.BC
has executed. the HL register pair will contain 249616.
034A
0000001101001010
214C = 0010000101001100
0010010010010110
No carry. set C to

1----; mmmm 2
....- - - 1 mmmm 3

0.J

tJ....
"'lI,..-

NO carry. set AC to 0
Addition instruction. set N to 0

The ADD HL.HL instructIOn

IS

equivalent to a 16-bit left shift.

3-50

ADD xy,rp-ADD REGISTER PAIR TO INDEX REGISTER

S

Z AC PIO N C

Data

FI \ Ixl \o\xl
A

S.C
D.E
H.l

rr

55

SP
PC

mmmm

IX

ppqq

IY
I
R

~ :...;;,mmm

.........

........

+')

" ~pqq+rr5~
...... -

Program
Memorv

l1vlll0l
OOxxl00l

mmmm
mmmm + 1

t-_ _-Immmm + 2
I -_ _-Immmm + 3

The illustration shows execution of ADD IX. DE.

/R
11..4- 11 01 00J- 1001

---J

o for Index register=IX

1 for Index reglster=IY

00
01
10
11

for
for
for
for

rp IS register pair BC
rp is register pair DE
rp IS specified Index register
rp is Stack POinter

Add the contents of the specified register pair to the contents of the specified Index
register.
Suppose IY contains 4FF016 and BC contains 000F16 After the instruction
ADD IY.BC
has executed. Index Register IY will contain 4FFF16

3-51

AND data-AND IMMEDIATE WITH ACCUMULATOR

S

Z AC PIO N

Data

F'xlxl lJxlo 10'

xx

A
B,C
D,E
H,L

SP
PC

_-~mmm+v

mmmm

IX
IY
I
R

z XX.VY.r
-

I-.

Program
Memorv

E6
yy

mmmm
mmmm+ 1
mmmm+2
I - - - - - - i mmmm + 3

AND

-.-

data

E6

yy

-.-

AND the contents of the next program memory byte to the Accumulator.
Suppose xx=3A 16. After the Instruction
AND 7CH
has executed, the Accumulator will contain 3816.
3A
7C

sets S to

001 1
01 1 1

1010
1 100

0011

1000

o..-J

LThree 1 bits, set PIO to 0

LNon-zero result. set Z to 0

This IS a routine logical instructIOn: It IS often used to turn bits "off" For example, the
Instruction
AND 7FH
will unconditionally set the high order Accumulator bit to O.

3-52

AND reg-AND REGISTER WITH ACCUMULATOR

S

"r,
} ,~Io'

Z ACP/O N C

xxyy

0 xl 1 X lOIoi
J

xx

B.C

--liIooA,B,C,D.E,
HorLlsyy

D.E
H.L

SP
PC
IX
IY
I

mmmm

:;;&-::1. mmmm + 1

......

I
I

Data

~
Program
Memorv

10100xxx

mmmm

I-_ _~mmmm + 1
I-_ _-Immmm + 2
I-_ _-Immmm + 3
AND

-..10100

reg
xxx

000

for reg=B
for reg=C
forreg=D
for reg=E
for reg=H
for reg=L
for reg=A

001
010
011
100
101
111

AND the Accumulator with the contents of ReglsterA. B, C. D. E. H or L. Save the resu It
in the Accumulator.
Suppose xx=E316' and Register E contains A016. After the instruction
AND E
has executed. the Accumulator will contain A016.
E3
AO =

1 1 10
10 10
1010

001 1
0000
0000

sets S to 1..-J

Two 1 bits. set P10 to 1

LNon-zero result. set Z to 0

AND is a frequently used logical instruction.

3-53

AND (HL) -AND MEMORY WITH ACCUMULATOR

AND (lX+disp)
AND (IV +disp)
S
F

Data
Memory

Z AC PIC N C

'XIXI11xloioi

__--f xx vv ) . . . . -

xx

A
B,C
D,E
H,L

~::J.mmmm+3

mmmm

--

ppqq

IY
I
R

ppqq +d

1\

SP
PC
IX

VV

I
I

ppqq+dJ-

Program
Memory

FD
A6
d

t----'==------f

mmm m
mmmm+l
mmmm+2
mmmm + 3

The illustration shows execution of AND (IY+disp),

AND (IY+disp)

~~

FD

A6

AND the contents of memory location (specified by the sum of the contents of the IY
register and.the displacement digit d) with the Accumulator.
Suppose xx=E316, ppqq=400016, and memory locatIOn 400F16 contains A016' After
the Instruction
AND (IY+OFH)
has executed, the Accumulator will contain A016'
E3

1 1 10

AO =

_~...:.~_~...:~_-,--~..:..~_~_~

1 sets S to

01 1 1

1~

LTWO 1 bitS, set PIO to 1

LNon-zero result. set Z to 0
AND (IX+disp)

~-.-

DO

A6

This instruction IS Identical to AND (IY+disp), except that It uses the IX register Instead
of the IY register,
AND (HL)
~

A6
AND the contents of the memory location (specified by the contents of the HL register
pair) with the Accumulator.
AND is a frequently used logical Instruction,

3-54

BIT b,reg -

s
F

TEST BIT b IN REGISTER reg

Z ACP/O N C

lulblllulol I

"

~
b

Data

A
B.C

yyy'Dyyyy

D.E

--...... --

H.L

SP
PC

mmmm

IX
IY
I
R

__~:::l mmmm+2

Program
Memory
CB
01bbbxxx

mmmm
mmmm + 1
mmmm+2
t - - - - I mmmm + 3

BIT
CB 01

b.

-.-

reg

~ ~

Bit Tested

a
1
2

3
4

5
6
7

Register

000
001
010
011
100
101
110
111

000
001
010
011
100
101
111

--BC

0
E
H
L

Place complement of Indicated register's specified bit in Z flag of F register.

Suppose Register C contains 1110 1111. The instruction BIT 4.C will then set the Z flag
to 1. while bit 4 In Register Cremains O. Bit a IS the least significant bit.

3-55

BIT b,(HL) - TEST BIT b OF INDICATED MEMORY POSITION

BIT b,UX+disp)
BIT b,(lY+disp)

ZACP/ON

Data
Memory

Flu!SlllulOI I

B.C
D.E

pp

../

yyy'Dyyyy

Program

p.q

qq

H.L

-,.

SP
PC
IX

mmmm

I -~:I mmmm + 2

Memory

IY
I

mm mm
CB
01bbb110 mm mm+1
mmmm+2
11----1 mmmm + 3

The illustration shows execution of BIT 4.(HL). Bit 0 IS the least significant bit.
BIT
-.-

b,

(HL)

CB 01

bbb
-.-

110

Bit Tested bbb

000
0
1
001
2
010
3
011
4
100
101
5
110
6
7
111
Test Indicated bit within memory position specified by the contents of Register HL. and
place bit's complement in Z flag of the F register.
Suppose HL contains 4000H and bit 3 In memory location 4000H contains 1. The instruction
BIT 3,(HL)
will then set the Z flag to 0, while bit 3 In memory location 4000H remains 1.
BIT b,(IX+disp)

~-

DD CB dOl bbb 110

bbb IS the same as in BIT b, (HL)

Examine specified bit within memory location Indicated by the sum of Index Register IX
and disp. Place the complement In the Z flag of the F register.

3-56

Suppose Index Reglste( IX contains 4000H and bit 4 of memory location 4004H IS 0,
The instruction
BIT 4, (IX +4H)
will then set the Z flag to 1, while bit 4 of memory location 4004H remains 0,

JK

FD CB d 01 bbb 110
bbb IS the same as In BIT b, (HL)

This instruction is identical to BIT b,(lX+displ. except that It uses the IY register instead
of the IX register,

CAll label S

CAll THE SUBROUTINE IDENTIFIED IN THE

OPERAND

Z AC Pia N C

! ! !

Data

I I I

Memorv

A
B,C

mm+3
mm

..I::xxxx-Z

D,E
H,L

SP
PC

../y
rAI"

xxxx
mmmm

IX

IY

Ii
R

b-~mmm+D-

xxxx-Z
xxxx-1
xxxx

Program
Memorv

mmmm
mmmm+ 1
{
qq
mmmm+Z
t--'-'--I mmmm + 3
CO

pp

CALL

label

"-v-'

"-v-'

CD

ppqq

Store the address of the instruction following the CALL on the top of the stack: the top
of the stack is a data memory byte addressed by the Stack POinter, Then subtract 2
from the Stack POinter in order to address the new top of stack, Move the 16-bit address
contained in the second and third CALL instruction object program bytes to the Program Counter. The second byte of the CALL instruction IS the low-order half of the address, and the third byte is the high-order byte,
ConSider the Instruction sequence:
CALL
AND

SUBR
7CH

SUBR
After the instruction has executed, the address of the AND instruction is saved at the
top of the stack. The Stack POinter IS decremented by 2, The instruction labeled SUBR
will be executed next.

3-57

CALL condition. label -

CALL THE SUBROUTINE IDENTIFIED IN

THE OPERAND IF CONDITION IS
SATISFIED

CALL

condition.

label

T~I
T
--11 xxx 100

000
001
010
011
100
101
110
111

pp qq

Condition
NZ Non-Zero
Z
Zero
NC Non-Carry
C Carry
PO Panty Odd
PE Parity Even
P Sign Positive
M Sign Negative

Relevant Flag
Z
Z

C
C
P/O
P/O
S
S

This instruction IS identical to the CALL instructIOn. except that the Identified
subroutine will be called onlv if the condition is satisfied: otherwise. the Instruction sequentially following the CALL condition instruction will be executed.
Consider the instruction sequence:
CALL

: COND.SUBR

_------1 condition not satisfied

AND

7CH

condition
satisfied

SUBR
If the condition IS not satisfied. the AND instruction will be executed after the CALL
COND.SUBR instruction has executed. If the condition IS satisfied. the address of the
AND instruction is saved at the top of the stack. and the Stack POinter IS decremented
by 2. The instruction labeled SUBR will be executed next.

3-58

CCF - COMPLEMENT CARRY FLAG

S

ZACP/ON C

[I::Ii:::II::I::J::J12X~!t:;;.i
. i=======~"'~~

A
B.C
D,E
H.L
SP
PC

mmmrn

-/ :v

_--l. mmmm+ 1

IX
IY
I
R

Data

~
Program

emorv

~
3F'

CCF
3F

Complement the Carry flag. No other status or register contents are affected,

3-59

mmmm
mmmm+l
mmmm+2
mmmm+3

CP data-COMPARE IMMEDIATE DATA WITH

ACCUMULATOR

S Z AC P/Q N C

Data

(XW]JxDW

r---<

xx

A
S,C

D,E
H,L
SP

PC

___
~mmm+.D

mmmm

IX
IY
I
R

xx-yv:r-

~
Program
Memorv

FE

mmmm

vy

mmmm+l
mmmm+2

mmmm+3
CP

data

FE

yy

Subtract the contents of the second oblect code byte from the contents of the Accumulator. treating both numbers as simple binary data. Discard the result i.e.. leave
the Accumulator alone, but modify the status flags to reflect the result of the subtraction.
Suppose xx=E316 and the second byte of the CP Instruction object code contains
AO 16. After the Instruction
CP OAOH
has executed, the Accumulator will still contain E316, but statuses will be modified as
follows:
1 1 10
1010
0100

E3
AO

o sets

001 1
0000
0011

fJ

S to 0

No borrow. set C to 0

re' Z" 0

'--- - - - - N o borrow, set AC to 0

1"+ 1=0, set P/O to 0

Notice that the resulting carry

LN,o'M' rewlt.

Subtract instruction. set N to 1

IS

complemented,

3-60

CP reg-COMPARE REGISTER WITH ACCUMULATOR

s
F

,,-

Z AC P/O N C "-

XX-YY " )

IxlxlxlxlJixl
1

xx

}-...- A,B,C,D,E,H
eoo.m,,'

B,C
D,E
H,l

SP
PC

or Lis yy

mmmm

~--::I

mmmm + 1

......

IX

Data

~
Program
Memory

10111xxx

..-----1
1-_ _--1
..-_ _--1

IY

CP

mmmm
mmmm + 1
mmmm + 2
mmmm + 3

reg

-.-

-.-

10111

xxx

(j'Q'Q
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Subtract the contents of Register A, B, C, D, E, H or L from the contents of the Accumulator. treating both numbers as simple binary data, Discard the result: i.e" leave
the Accumulator alone, but modify status flags to reflect the result of the subtraction,
Suppose xx=E316 and Register B contains A016' After the Instruction
CP B
has executed, the Accumulator will still contain E316, but statuses will be modified as
follows:
E3 =

1 1 10

0011
0000

AO = 1 0 1 0

fJ

101 00

0"" S 10 0
No borrow, set C to 0

001 1

Noo .""

"""'t. '"

Z 10 0

No borrow. set AC to 0
Subtract instruction, set N to 1

1 1=0, set P/O to 0

Notice that the resulting carry

IS

complemented,

3-61

CP (Hl) - COMPARE MEMORY WITH ACCUMULATOR

CP (lX+disp)
CP (lY +disp)
,
,

~
Ixlxlxlxlllx'

Data
Memory

---< -

"-

xx

XX-yy.r-

B.C
D.E
H.L

SP
PC

qq

PP

~mmm+~

mmmm

IX
IY

The illustration shows execution of CP (HL!:

CP (HL!

P qq

Program
Memory

~
BE

yy

mmm

mmmm+l
mmmm+2
mmmm+3

BE
Subtract the contents of memory location (specified by the contents of the HL register
pair) from the contents of the Accumulator. treating both numbers as simple binary
data. Discard the result: I.e.. leave the Accumulator alone. but modify status flags to
reflect the result of the subtraction.
Suppose xx=E316 and yy=A016. After executIOn of
CP (HU
the Accumulator will still contain E316. but statuses will be modified as follows:

E3
AO =

0"" S ,,0
No borrow. set C to 0

1 1 10
0 1 10

001 1
0000

0100

0011

fJ

LNoo.w",,"". '" '<00

No borrow. set AC to-O

1 1=0. set P10 to 0

Notice that the resulting carry

Subtract instruction. set N to 1

IS

complemented.
CP (IX+disp)
-...--,-

DO BE

3-62

Subtract the contents of memory location (specified by the sum of the contents of the
IX register and the displacement value d) from. the contents of the Accumulator. treatIng both numbers as simple binary data. Discard the result I.e., leave the Accumulator
alone. but modify status flags to reflect the result of the subtraction.
CP (lY+disp)
--...-.FD BE

This instruction IS Identical to CP (IX +disp), except that It uses the IY register Instead of
the IX register.

CPO - COMPARE ACCUMULATOR WITH MEMORY.

DECREMENT ADDRESS AND BYTE COUNTER

t
[iJx Ix r{!1
5

...J..
' - xx-yy

Z ACP/ON C'

Set if BC-1 ';'0.

reset otherwise

B.C
D.E
H.L
SP
PC
IX

xx
uu

tt
pp

qq

mmmm

Data
Memory

lluu-1 ' )

"/

"I

"'*"

VV

.-rrf' ppqq-1

t-....

IV

mmmm+v

I
I

Program
Memory
ED
A9

pp qq

mmmm
m mmm+l

I -_ _~ mmmm + 2
1 - - - - 1 mmmm + 3

CPO
'-v-"

ED A9

Compare the contents of the Accumulator with the contents of memory location
(specified by the HL register pair), If A IS equal to memory. set Z flag. Decrement the HL
and BC register pairs. (BC IS used as the Byte Counter.)

3-63

Suppose xx=E316. ppqq=400016. BC contains 000116. and yy=A016 After the Instruction
CPO
has executed. the Accumulator will still contain E316. but statuses will be modified as
follows:
E3
AO

1 1 10
1010
0100

001 1
0000
0011

oOO~ o.J }.J

S"

LN,"~"" ""It.

001 Z "

- - - - N o borrow. set AC to 0

l,-_

The P10 flag will be reset

because BC-1 =0
Subtract Instruction involved.
set N to 1
Carry not affected.
The HL register pair will contain 3FFF16. and BC=O.

CPDR-COMPARE ACCUMULATOR WITH MEMORY.

DECREMENT ADDRESS AND BYTE COUNTER.
CONTINUE UNTIL MATCH IS FOUND OR BYTE
COUNTER IS ZERO
CPDR
~

ED B9
ThiS instruction IS Identical to CPO. except that it IS repeated until a match is found or
the byte counter is zero. After each data transfer. Interrupts will be recognized and two
refresh cycles will be executed.
Suppose the HL register pair contains 500016. the BC register pair contains 00FF16.
the Accumulator contains F916. and memory has contents as follows:
Location

Contents

500016
4FFF16
4FFE16
4FF016
4FFC16
4FFB16

AA16
BC16
1916
7A16
F916
0016

After execution of
CPDR
the P/O flag will be 1. the Z flag will be 1. the HL register pair will contain 4FFB16. and
the BC register pair will contain 00FA16.

3-64

CPI-COMPARE ACCUMULATOR WITH MEMORY.

DECREMENT BYTE COUNTER.
INCREMENT ADDRESS

A
B,C
D,E
H,L

.J.
xx-yy

Z ACP/ON C

eX IXI x I \ : ! J ereset
t if BC-l ,. 0,
otherwise

L -

pp

Data
Memorv

tluu-l

xx
uu

tl

"/

yv

I.tIII"

r;..a--

qq

ppq q

Ppqq'+l)

SP
PC
IX

IY
I
R

mmmm

Program
Memorv

"'-'ltiI..

~mmm+v

I
I

m mmm
m mmm+l
mmmm+2
t - - - - I mmmm + 3
ED
Al

CPI

-..-ED A1
Compare the contents of the Accumulator with the contents of memory location
(specified by the HL register pair!, If A IS equal to memory, set the Z flag, Increment the
HL register pair and decrement the Be register pair (BC IS used as Byte Counter!,
Suppose xx=E316, ppqq=400016, BC contains 003216, and yy=E316' After the instruction
CPI
has executed, the Accumulator will still contain E316' but statuses will be modified as
follows:
E3
-E3

0"" S"

1 111
0000

001 1
1 10 1

0000

0000

o---l

LA,,,,,,, o. '" ",1

~,----No

borrow, set AC to 0

The P/O flag will be set

because BC-1 oF 0,
Subtract Instruction Involved,
set N to 1.
Carry not affected.
The HL register pair will contain 400116, and BC will contain 003116.

3-65

CPIR-COMPARE ACCUMULATOR WITH MEMORY.

DECREMENT BYT.E COUNTER.
INCREMENT ADDRESS.
CONTINUE UNTIL MATCH IS FOUND
OR BYTE COUNTER IS ZERO
CPIR

-..-..ED B1
This instruction IS Identical to CPI. except that it IS repeated until a match IS found or
the byte counter is zero. After each data transfer interrupts will be recognized and two
refresh cycles will be executed.
Suppose the HL register pair contains 450016. the BC register pair contains 00FF16.
the Accumulator contains F916. and memory has contents as follows:
Location
-450016
4501 16
450216

Contents
AA16
15 16
F916

---

After execution of

CPIR
the P/0 flag will be 1. and the Z flag will be 1. The HL register pair will contain 450316.
and the BC register pair will contain 00FC16.

3-66

CPL- COMPLEMENT THE ACCUMULATOR

S

ZACP/ON

Data

FOJI[]:[O
-~

....... xx

xx

::::""""1

B,C

D,E
H,L

SP
PC
IX
IY

I
R

~"""'mmmm+)

mmmm

CPL

~
Program
emorv

~
2F

mmmm
mmmm+ 1
mmmm+2
mmmm+3

2F
Complement the contents of the Accumulator. No other register's contents are
affected.
Suppose the Accumulator contains 3A16 After the instruction

CPL
has executed, the Accumulator will contain C516
3A = 001 1
Complement = 1 1 00

1010
0101

This IS a routine logical Instruction. You need not use it for binary subtraction: there are
special subtract instructions (SUB, SBC).

3-67

DAA S

DECIMAL ADJUST ACCUMULATOR

ZACP/ON

Data

nXlxlxlxl Ixl
xx

t:)

I.-..

.!"Cpnvert
_deCimal

I.-..

~mmm+~

B.C

D.E
H.L

SP
PC

mmmm

IX
IY
I

Program
emorY

~
27

DAA

mmmm
mmmm+1
mmmm+2
mmmm+3

27
Convert the contents of the Accumulator to binary-coded decimal form. This instruction should only be used after adding or subtracting two BCD numbers: i.e.. look upon
ADD DAA or ADC DAA or INC DAA or SUB DAA or SBC DAA or DEC DAA or NEG DAA
as compound. decimal arithmetic Instructions which operate on BCD sources to generate BCD answers.
Suppose the Accumulator contains 3916 and the B register contains 4716. After the instructions
ADD B
DAA
have executed. the Accumulator will contain 8616. not 8016.

l80 CPU logic uses the values In the Carry and Auxiliary Carry. as well as the Accumulator contents. In the Decimal Adjust operation.

3-68

DEC reg 5

DECREMENT REGISTER CONTENTS

Z ACP/O N C

Data

F[[IXIX I x 11 LI

}L

A
B.C

D.E
H.l
SP
PC
IX

mmmm

:.:-

C,,""U of A
B. C. D. E. H.
orL,syy

:I~mmm ~
+

~
Program
Memorv

IY

I
R

OOxxxl0l

i
i

mmmm

I-_ _-Immmm + 1
I-_ _-Immmm + 2
1-_ _-1 mmmm + 3

R
00

xxx

101

000
001
010
011
100
101
111

for reg=B
for reg=C
forreg=D
for reg=E
for reg=H
forreg=L
for reg=A

-,-

Subtract 1 from the contents of the specified register.

Suppose Register A contains 5016. After execution of
DEC A
Register A will contain 4F16.

3-69

DEC rp DEC IX
DECIY
S

DECREMENT CONTENTS OF SPECIFIED REGISTER

PAIR

Z AC PIO N C

I ! !

Data

! I

S.C
D.E

~COO,,""
DE. HL or"SP"

H.L

IS YYYY

SP
PC
IX

mmmm

_--~mmm+v

~
Program
Memorv

IY

OOxxl0ll

mmmm
mmmm+l
t - - - - ; mmmm + 2

t-----I mmmm + 3
The illustratIOn shows execution of DEC rp:

1l

00 xx 1011

-.00
01
10
11

for
for
for
for

rp
rp
rp
rp

IS
IS
IS
IS

register pair BC
register pair DE
register pair HL
Stack POinter

Subtract 1 from the 16-blt value contained In the specified register pair. No status flags
are affected.
Suppose the Hand L registers contain 2F0016. After the instruction
DEC HL
has executed. the Hand L registers will contain 2EFF16.
DEC IX

--..-

DD 2B
Subtract 1 from the 16-blt value contained In the IX register.
DEC IY

--..FD 2B

Subtract 1 from the 16-blt value contained In the IY register.

Neither DEC rp. DEC IX nor DEC IY affects any of the status flags. ThiS IS a defect in the
Z80 instruction set. Inhented from the 8080. Whereas the DEC reg instruction is used In
Iterative instruction loops that use a counter with a value of 256 or less. the DEC rp
(DEC IX or DEC IY) instruction must be used if the counter value is more than 256. Since
the DEC rp Instruction sets no status flags. other instructions must be added to simply

3-70

test for a zero resu It. This

IS

a typical loop form:

LD

DE. DATA

:LOAD INITIAL 16-81T COUNTER VALUE

:FIRST INSTRUCTION OF LOOP

DEC
LD
OR

DE
A.D

JP

NZ.LOOP

:DECREMENT COUNTER
TO TEST FOR ZERO. MOVE D TO A
:THEN OR A WITH E
:RETURN IF NOT ZERO

LOOP

DEC (HLl - DECREMENT MEMORY CONTENTS

DEC (IX+disp)
DEC (lY+disp)
S Z AC P/O N

Data

F([lxjxJXJ1!
yy-l

P;

VV

S.C

D.E
pp

H.L

qq

SP
PC

mmmm

IX
IV
I

;;;&'""::;;'mmm +

Program

Memory

~
35

The illustration shows execution of DEC (HL):

mmm

m mmm+l
mmmm+2
mmmm+3

DEC (HL)
'-v-"

35
Subtract 1 from the contents of memory location (specified by the contents of the HL
register paid.
Suppose ppqq=450016. yy=5F16' After execution of
DEC (HL)
memory location 450016 will contain 5E16.

__

5F = 0 1 0 1
-01 = 1 111

1111
1 111

10101

1 1 10

0 "" S to O.JJ

1 1=0. set P10 to 0

r
-

L Noo."" ",olt. '" Z

to 0

No borrow, set AC to 0
Subtract instruction, set N to 1

3-71

DEC (IX+disp)

~-.-

DO 35

Subtract 1 from the contents of memory location (specified by the sum of the contents
of the IX register and the displacement value dl.
DEC (IY+disp)

~-,-

FD 35

This instruction is Identical to DEC (IX+displ. except that it uses the IY register instead
of the IX register.

01 -

DISABLE INTERRUPTS
S

Z AC Pia N C

Data

FI I i I I I I
A

S.C
D.E
H.L

SP
PC
IX
IY
I
R

mmmm

,. -

t-.~

......mmmm+l

Program
emorv

~
F3

01

mmmm

mmmm+l
mmmm+2
mmmm+3

F3
When this instruction is executed. the maskable interrupt request IS disabled and the
INT input to the CPU will be Ignored. Remember that when an interrupt is
acknowledged. the maskable interrupt IS automatically disabled.
The maskable interrupt request remainS disabled until It IS subsequently enabled by an
EI instruction.
No registers or flags are affected by this instruction.

3-72

DJNZ disp -

JUMP RELATIVE TO PRESENT

CONTENTS OF PROGRAM COUNTER IF
REG B IS NOT ZERO

ZACP/ON

I I I I I I
A
S.C
D.E
H.L

xx-1

Data

T..

xx

SP
PC
IX
IY
I

........mmmm+
I-.~:::l {dd-2l + 2

mmmm

---DJNZ

disp

10

dd-2

~
Program

Memorv

10
dd-2

mmmm
mmmm+ 1
mmmm+2
mmmm+3

Decrement Register B. If remaining contents are not zero. add the contents of the DJNZ
Instruction object code second byte and 2 to the Program Counter. The lump is
measured from the address of the Instruction operation code. and has a range of -126 to
+ 129 bytes. The Assembler automatically adjusts for the tWice-Incremented PC.
If the contents of B are zero after decrementing. the next sequential Instruction IS executed.
The DJNZ instruction IS extremely useful for any program loop operation. since the one
Instruction replaces the typical "decrement-then-branch on condition" instruction sequence.

EI- ENABLE INTERRUPTS

S

Z AC PIO N C

Data

diLl! []
A

B.C
D.E
H.L

.-mmmm+ 1

SP
PC
IX
IY

mmmm

~~::I

"-

Program
emorv

~
FB

3-73

mmmm
mmmm+ 1
mmmm+2
mmmm+3

EI
FB
Execution of this instruction causes interrupts to be enabled. but not until one more instruction executes.
Most Interrupt service routines end with the two instructions:
EI
RET

:ENABLE INTERRUPTS
:RETURN TO INTERRUPTED PROGRAM

If Interrupts are processed serially. then for the entire duration of the interrupt service
routine all maskable interrupts are disabled - which means that In a multi-interrupt
application there IS a Significant possibility for one or more Interrupts to be pending
when any Interrupt service routine completes execution.
If Interrupts were acknowledged as soon as the EI instructions had executed, then the
Return Instruction would not be executed. Under these circumstances. returns would
stack up one on top of the other - and unnecessarily consume stack memory space.
This may be illustrated as follows:
Interrupt

Interrupt service routine

By inhibiting Interrupts for one more instruction following execution of EI. the Z80 CPU
ensures that the RET instruction gets executed in the sequence:

EI
RET

:ENABLE INTERRUPTS
;RETURN FROM INTERRUPT

It IS not uncommon for interrupts to be kept disabled while an interrupt service routine
is executing. Interrupts are processed serlallv:

7'-\- 7'"'"'"\ Interrupt service routine

Interrupt service routine

3-74

EX AF,AF' -EXCHANGE PROGRAM STATUS AND ALTERNATE

PROGRAM STATUS
S

Z ACP/ON C

Alternate
Register Set

D.E

H.L

SP
PC

mmmm

IX
IY
I
R

F'
A'
B',C'

A
B.C

_~-t'mmmm + ~

EX AF.AF
~

H',L'

Program

~
emorY

08

D',E'

mmmm
mmmm+ 1
mmmm+2
mmmm+3

08
The two-byte contents of register pairs AF and AT are exchanged.
Suppose AF contains 4F9916 and AT contains lOAA 16 After execution of
EX AF,AF
AF will contain lOAA16 and AF will contain 4F9916.

3-75

EX DE,HL -

EXCHANGE DE AND HL CONTENTS

SZACP/ONC

FO:I

Data

10

S,C
D,E
H,L
SP
PC
IX
IY
I

pp

qq

;..:

xx

VV

mmmm

"--~mmm+~

~
Program

emorv

~
ES

EX DE,HL
~

mmmm
mmmm+l
mmmm+2
mmmm+3

EB

The D and E registers' contents are swapped with the Hand L registers' contents,
Suppose pp=0316, qq=2A 16, xx=4116 and yv=FC 16, After the instructIOn
EX DE,HL

has executed, H will contain 0316, L will contain 2A 16, 0 will contain 4116 and E will
contain FC 16,
The two Instructions:
EX DE,HL
LD A,(HL)

are equivalent to:

LD A,(DE)

but if you want to load data addressed by the D and E register into the B register,
EX DE,HL
LD B,(HL)

has no single instruction equivalent.

3-76

EX (SP) ,HL EX (SP),IX

EX (SP),IY
S

FI

EXCHANGE CONTENTS OF REGISTER AND

TOP OF STACK

ZACP/ON C

Data
Memory

I I I I I I

qq

B.C

pp

D.E
H.L

xx

yy

SP

ssss
mmmm

PC
IX

_~

IY

I
I

I
R

mmmm+l

Program
Memory

~
E3

The illustration shows execution of EX (SP).HL.

ssss
ssss+ 1
ssss + 2

mmmm

mmmm+l
mmmm+2
mmmm+3

EX (SP).HL
~

E3
Exchange the contents of the L register with the top stack byte. Exchange the contents
of the H register with the byte below the stack top.
Suppose xx=2116. yv=FA16. pp=3A16. qq=E216. After the Instruction
EX (SP).HL
has executed. H will contain 3A 16. L will contain E216 and the two top stack bytes will
contain FA16 and 2116 respectively.
The EX (SPl.HL Instruction is used to access and manipulate data at the top of the stack.
EX (SPl.IX
~

DD E3
Exchange the contents of the IX register's low-order byte with the top stack bvte. Exchange the IX register's high-order byte with the byte below the stack top.
EX (SPl.IY
~

FD E3
This Instruction
register.

IS

identical to EX (SPl.IX. but uses the IY register Instead of the IX

3-77

EXX -

EXCHANGE REGISTER PAIRS AND ALTERNATE

REGISTER PAIRS
Z AC PIG N C

FI I I

Altemate
Regtster Set

I I I

.. ffi ~"'.C'
F'

}.

B,C
D,E
H,L

SP

PC

mmmm

__~fmmmm+v

IX
IY
I

D',E'
H',L'

Program
emorv

~
.

EXX

D9

mmmm
mmmm+l
mmmm+2
mmmm+3

09
The contents of register pairs BC, DE and HL are swapped with the contents of register
pairs B'C', D'E', and H'L'
Suppose register pairs BC, DE and HL contain 490116, 5F0016 and 725116 respectively, and register pairs B'C', D'E', H'L' contain 000016, lOFF16 and 333316 respectively. After the execution of

EXX
the registers will have the following contents:

BC: 000016: DE. lOFF16: HL: 333316:

B'C' 490116: D'E' 5F0016: H'L' 725116
This instruction can be used to exchange register banks to provide very fast Interrupt
response times,

3-78

HALT
S

FI I

ZACP/ONC

Data

B,C
D,E
H,L
SP
PC

mmmm

~mmm+~

.....

IX
IY
I
R

Program
emorv

~
76

HALT

mmmm
mmmm+ 1
mmmm+2
mmmm+3

76
When the HALT instruction IS executed, program execution ceases, The CPU requires
an interrupt or a reset to restart execution, No registers or statuses are affected:
however, memory refresh logic continues to operate,

3-79

1M 0 -INTERRUPT MODE 0
5
F

A
S.C
D.E
H.L
SP
PC
IX

Z ACP/0N

Data

I I I I I I I

mmmm

IY
I

-- ~

L-..... ::I mmmm + 2

Program
Memory

ED
46

mmmm
mmmm+ 1

I - - - - - i mmmm + 2
1------lI mmmm + 3

--.--IMO

ED 46

This instruction places the CPU In interrupt mode O. In this mode. the interrupting
deVice will place an instructIOn on the Data Bus and the CPU will then execute that instruction. No registers or statuses are affected.

1M 1 -INTERRUPT MODE 1
IM1
--.--ED 56
This instruction places the CPU In Interr.upt mode 1. In this mode. the CPU responds to
an Interrupt by executing a restart (RST) to location 003816.

1M 2 -INTERRUPT MODE 2
1M 2

--.,ED 5E

This instruction places the CPU In interrupt mode 2. In this mode. the CPU performs an
indirect call to any specified location In memory. A 16-bit address IS formed uSing the
contents of the Interrupt Vector (I) register for the upper eight bits. while the lower
eight bits are supplied by the interrupting device. Refer to Chapter 12 for a full descnption of interrupt modes. No registers or statuses are affected by this instruction.

3-80

IN A.(port) -INPUT TO ACCUMULATOR

S

Z AC P/O N C

I I I I I I I

Data

14-

A
B.C
D.E
H.l

SP
PC

I/O port yy

mmmm

I-.

IX

-=<:mmm +,V

~
Program
Memory

IV
I

DB

mmmm

YV

mmmm+ 1
mmmm+2
mmmm+3

--IN A.

DB

(port)
'-v-'

yy

Load a byte of data Into the Accumulator from the I/O port (identified by the second IN
instruction object code byte).
Suppose 3616 is held In the buffer of I/O port 1A 16. After the instruCtion
IN A.(lAH)
has executed. the Accumulator will contain 3616.
The IN Instruction does not affect any statuses.
Use of the IN Instruction IS very hardware dependent. Valid I/O port addresses are
determined by the way in which I/O logic has been Implemented. It IS also possible to
deSign a microcomputer system that accesses external logiC uSing memory reference
instructions with specific memory addresses.

3-81

INC reg -INCREMENT REGISTER CONTENTS

S

Z AC

plo

N C

Flxlxlxlxlol I

~e-"o..or

B.C

B. C. D. E. H

DE

LIs YV

_/ - V

H.l

SP
PC

mmmm

_.--3.. mmmm +

IX
IY

I
R

Program
Memorv

OOxxx 100 mmmm

mmmm+l
1----1 mmmm + 2

1----1 mmmm + 3

it
00 xxx 100
000
001
010
all
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg =H
reg=L
reg=A

Add 1 to the contents of the specified register.

Suppose Register E contains A816. After execution of
INC E
Register E will contain A916.

3-82

INC rp -INCREMENT CONTENTS OF SPECIFIED REGISTER PAIR

INC IX
INC IV
S Z AC PIO N C

F[!

Data

n ] []

A
B.C
D,E
H,L
SP
PC

Co~mOC
DE. HL or SP

--'15 YYYY

"
......

:;;a:-:I mmmm + 1

mmmm

IX

IY
I
R

I
I

~
Program

Memory

OOxxOO11

mmmm

f-_ _-Immmm + 1
mmmm+2

f - - - - I mmmm + 3
The illustration shows execution of INC rp:

~
00 xx 0011
00
01
10
11
Add 1 to the 16-bit value contained
affected.

In

for
for
for
for

rp
rp
rp
rp

IS
IS
IS
IS

register pair BC
register pair DE
register pair HL
Stack POinter

the specified register pair. No status flags are

Suppose the D and E registers contain 2F7 A 16. After the instruction
INC DE

---

has executed. the D and E registers will contain 2F7B16.

INC IX

DD 23
Add 1 to the 16-blt value contained

In

the IX register.
INC IY

FD 23

Add 1 to the 16-blt value contained in the IY register.

Just like the DEC rp. DEC IX and DEC IY, neither INC rp. INC IX nor INC IY affects any
status flags. This is a defect In the Z80 instruction set inherited from the 8080.

3-83

INC (HL) -INCREMENT MEMORY CONTENTS

INC (IX +disp)
INC (lY+disp)
S

Z AC Pia N

'xixlxlxioi I
A
B.C
D.E
H.L

PC

mmmm

IX
IV

ppqq

I
R

I:!.-

yy+l

.. /"

SP

Data
Memory

~ -::l mmmm + 3

~pqq+d~

yy

./

Program
Memory
DD
34
d

pp qq

+d

t
m mmm
m mmm+ 1
m mmm+2
m mmm+3

The illustration shows execution of INC (IX+d):

INC (IX+disp)
"-v-'

-.-

DO 34

Add 1 to the contents of memory location (specified by the sum of the contents of
Register IX and the displacement value d).
Suppose ppqq=400016 and memory location 400F16 contains 3616. After execution
of the instruction
INC (IX+OFH)
memory location 400F16 will contain 3716.

36 = 001 1

0110

01011

o sets

...

S to 0

Carry status not affected

01 1 1

OO=O. set PIO to 0

LNoo.",o re"'t ," Z" 0

No carry. set AC to 0
Addition instruction. set N to 0

INC (IY+disp)
"-v-'

-.-

FD 34

This Instruction IS Identical to INC (IX+disp). except that it uses the IY register Instead
of the IX register.
INC (HL)
~

34
Add 1 to the contents of memory location (specified by the contents of the HL register
pair).

3-84

IND -INPUT TO MEMORY AND DECREMENT POINTER

Z ACP/O N C

lulxlulull!

,C

xx-l

I/O port yy

Memory

A
S.C
D.E

xx

H.l

pp

yy

I..-

mmmm

mmmm+2

IX
IY
I

ppqq

.....I'" ppqq-l

qq

SP
PC

Data

Program
Memory
mmm m
mmm m+ 1
mmmm+2
i - - - - - 1 mmmm + 3
ED

AA

INo

-.--

ED AA
Input from I/O port (addressed by Register C) to memory location (specified by HU.
Decrement Registers Band HL.
Suppose xx=0516. yy=1516. ppqq=240016. and 1916 IS held in the buffer of I/O port
1516. After the instruction
INo
has executed. memory location 240016 will contain 1916. The B register will contain
0416 and the HL register pair 23FF16

INDR -INPUT TO MEMORY AND DECREMENT POINTER

UNTIL BYTE COUNTER IS ZERO
INoR

-.--

ED BA

INoR IS Identical to IND. but IS repeated until Register B=O.

Suppose Register B contains 0316. Register C contains 1516. and HL contains 240016.
The following sequence of bytes IS available at I/O port 1516:
1716.5916 and AE16
After the execution of
INoR
the HL register pair will contain 23Fo16 and Register B will contain zero, and memory
locations will have contents as follows:
Location
2400
23FF
23FE

Contents
17 16
59 16
AE16

This instruction is extremely useful for loading blocks of data from an Input device Into
memory.

3-85

INI -

INPUT TO MEMORY AND INCREMENT POINTER

Z AC Pia N C

I I I I I I I

I/O port yy

D.E
H.L

SP
PC

xx

Data
Memory

IL__.....I===~

B.C

VV

t----::7'"---+---=.::----i-~~~
ppqq + 1
pp
qq

ppqq

~-

~mmm+v

mmmm

IX
IV

I
R

Program
Memory

ED
A2

mmmm
mmmm+ 1
mmmm+2

I - - - - - f mmmm + 3

INI

-..--

ED A2

Input from I/O port (addressed by Register CI to memory location (specified by HL).
Decrement Register B: Increment register pair HL
Suppose xx=0516. yy= 1516. ppqq=240016. and 1916 IS held In the buffer of I/O port
15 16.
After the instruction
INI
has executed. memory location 240016 will contain 1916. The B register will contain
0416 and the HL register pair 240116.

INIR -INPUT TO MEMORY AND INCREMENT POINTER

UNTIL BYTE COUNTER IS ZERO
INIR
-..-ED B2
INIR IS Identical to IN!. but IS repeated until Register B=O.
Suppose Register B contains 0316. Register C contains 1516. and HL contains 24016.
The following sequence of bytes IS available at I/O port 1516'
1716. 5916 and AE16
After the execution of
INIR
the HL register pair will contain 240316 and Register B will contain zero. and memory
locations will have contents as follows:
Location
2400
2401
2402

Contents
17 16
59 16
AE16

ThiS instruction IS extremely useful for loading blocks of data from a deVice Into memory.

3-86

IN reg,le) -INPUT TO REGISTER

S

Z ACP/O N C

Ixlxlolxlol I

.1

Register

S.C

A. B. C. D. E.

D.E

H or L

H.L

SP
PC

Data

I/O port yy I

mmmm

"........

__--::I mmmm + 2

IX
IY
I

Program
Memorv

ED

mmmm
mmmm + 1
mmmm+2
.....- - - - I mmmm + 3

01xxxOOO

KED 01 xxx 000

000
001
010
all
100
101
111
110

for reg=B
for reg=C
for reg=D
for reg=E
for reg=H
forreg=L
for reg=A
for setting of status flags without
changing registers

Load a byte of data Into the specified register (reg) from the I/O port (identified by the
contents of the C register).
Suppose 4216 is held In the buffer of I/O port 3616. and Register C contains 3616.
After the Instruction
IN D.(C)

has executed. the D register will contain 4216.

During the execution of the instruction. the contents of Register B are placed on the top
half of the Address Bus. making It possible to extend the number of addressable I/O
ports.

3-87

JP label S

JUMP TO THE INSTRUCTION IDENTIFIED

IN THE OPERAND

Z AC PIO N C

Data

""1_....."--'-_....1_'

B.C
D.E
H.L
SP
PC

mmmm

IX
IY

..-( ppqq)

Program
Memory

I
I

C3
qq
pp

mmmm
mmmm+ 1
mmmm+2
mmmm+3

JP label
-.'-.,,-'
C3

ppqq

Load the contents of the Jump instruction obJect code second and third bytes Into the
Program Counter: thiS becomes the memory address for the next instruction to be executed. The prevIous Program Counter contents are lost
In the follOWing sequence:
JP
AND

NEXT

NEXT
7FH

CPL

The CPL instruction will be executed after the JP instruction. The AND instruction will
never be executed. unless a Jump instruction somewhere else In the instruction sequence lumps to thiS instruction.

3-88

JP condition, label- JUMP TO ADDRESS IDENTIFIED IN THE

OPERAND IF CONDITION IS
SATISIFED
JP condo label

HI
11 cc 010 ppqq

000
001
010

NZ
Z
NC
C
PO
PE
P
M

all
100
101
110
111

Condition
Non-Zero
Zero
No Carry
Carry
Parrty Odd
Parrty Even
Sign Positive
Sign Negative

Relevant Flag
Z
Z
C
C

Pia
Pia
S
S

This Instruction IS Identical to the JP Instruction. except that the lump will be performed only if the condition IS satisfied: otherwise. the Instruction sequentially followIng the JP condition Instruction will be executed.
Consider the Instruction sequence

JP
condition

AND

satisfied

LABEL

OR

COND.LABEL
I condition not satisfied
7CH

After the JP cond.label Instruction has executed. if the condition IS satisfied then the
OR Instruction will be executed. If the condition IS not satisfied. the AND Instruction.
being the next sequential Instruction. IS executed.

3-89

JP (Hll JP (IX)
JP (IV)
S

JUMP TO ADDRESS SPECIFIED BV CONTENTS

OF 16-BIT REGISTER

Z ACPfON C

Data

FCIIIIIJ
A

B.C
D.E
H.L

qq

pp

SP
mmmm

PC

IX
IY
I
R

Program
emorv

~
E9

The Illustration shows execution of JP (HU:

mmmm
mmmm+1
mmmm+2
mmmm+3

JP (HU
'-v-"

E9
The contents of the HL register pair are moved to the Program Counter: therefore. an
Implied addressing lump IS performed.
The Instruction sequence
LD
JP

H.ADDR
(HU

has exactly the same net effect as the single InstructIOn

JP

AD DR

Both specify that the instruction with label ADDR IS to be executed next.
The JP (HU instruction IS usefu I when you want to Increment a return address for a
subroutine that has multiple returns.
Consider the following call to subroutine SUB:
CALL
JP

SUB
ERR

;CALL SUBROUTINE
;ERROR RETURN
:GOOD RETURN

Using RET to return from SUB would return execution of JP ERR: therefore. if SUB executes without detecting error conditions. return as follows;
POP

HL

:POP RETURN ADDRESS TO HL

INC
INC
INC
JP

HL
HL
HL
(HU

;ADD 3 TO RETURN ADDRESS

:RETURN
JP (IX)
-.--'

DD9
This instruction IS identical to the JP (HU instruction. except that It uses the IX register

3-90

Instead of the HL register pair.

JP (lY)

--.,FD E9

This instruction IS Identical to the JP (HU instruction, except that It uses the IY register
instead of the HL register pair,

JR C,disp-JUMP RELATIVE TO CONTENTS OF PROGRAM

COUNTER IF CARRY IS SET
JR C, disp
--.,--.-

38

dd-2

This instruction IS identical to the JR disp instruction, except that the lump IS only executed if the Carry status equals 1, otherwise, the next instruction IS executed,
In the following instruction sequence:
I

4000

JR

C,$+8

, . . - - - - - - - - - - 1 c=o

4002

AND

C=l
'----111.... 4008

OR

7FH

After the JR C,$+8 instruction, the OR instruction IS executed if the Carry status equals
1, The AND instruction IS executed if the Carry status equals 0,

3-91

JR disp S

JUMP RELATIVE TO PRESENT CONTENTS OF

PROGRAM COUNTER

Z ACP/0N

I ! I

Data

I I I

B.C
D.E

H.L

SP
PC
IX

rmmmm+
~ (dd-21+2

mmmm

IY
I
R

I
I

~
Program

Memory
18
dd-2

mmmm
mmmm+ 1
mmmm+2
mmmm+3

JR disp
-.--.18 dd-2
Add the contents of the JR instruction object code second byte. the contents of the Program Counter, and 2. Load the sum Into the Program Counter. The lump IS measured
from the address of the instructIOn operation code, and has a range of -126 to + 129
bytes. The Assembler automatically adiusts for the tWice-incremented PC.
The following assembly language statement
dress 400016.

IS

used to Jump four steps forward from ad-

JR $+4
Result of this instruction is shown below:
Location
4000
4001
4002
4003
4004

Instruction
18
02

......
111--- new PC value

3-92

JR NC,disp-JUMP RELATIVE TO CONTENTS OF PROGRAM

COUNTER IF CARRY FLAG IS RESET
JR NC.disp

-,.--.30

dd-2

r-_~4000

c=o

4001
4002
4003
4005

After the JR NC.$-3 instruction. the OR instruction IS executed if the Carry status equals
1. The ADD instruction IS executed if the Carry status equals O.

JR NZ,disp-JUMP RELATIVE TO CONTENTS OF PROGRAM

COUNTER IF ZERO FLAG IS RESET
JR NZ.disp

-,.--.20

dd-2

ThiS instruction is Identical to the JR disp instruction. except that the jump IS only executed if the Zero status equals 0: otherwise. the next Instruclion IS executed.
In the follOWing Instruclion sequence:
I

_ _ _...;4..;;.0.;;..00,,--_.;;..JR""--1'

z=o

:~~~

AND

NZ.$+6

+z=~FH

4005

'---"'4006

OR

After the JR NZ.$+6 instruction. the OR instruction IS executed if the Zero status equals

O. The AND instruction IS executed if the Zero status equals 1.

3-93

JR Z,disp -

JUMP RELATIVE TO CONTENTS OF PROGRAM

COUNTER IF ZERO FLAG IS SET
JR Z.disp
--.,..---.28

dd-2

This instruction IS Identical to the JR disp instruction. except that the lump IS only executed if the Zero status equals 1 otherwise. the next Instruction IS executed.
In the following Instruction sequence:

Z=l

I
l

4000

JR

4002

AND:

Z.$+6

7FH

4004
4005
4006

OR

t z=o

After the JR Z.$+6 instruction, the OR Instruction IS executed if the Zero status equals
1. The AND Instruction IS executed if the Zero status equals O.

LD A,I- MOVE CONTENTS OF INTERRUPT VECTOR OR

LD A,R REFRESH REGISTER TO ACCUMULATOR
S

Z AC PIO N

Data

F'xlxlolxlol
xx

B.C
D,E
H.l

SP

PC
IX
IY
I

__,...(""mmmm + )

mmmm

Program

Memory

xx

ED
57

mmmm
mmmm+ 1
mmmm+2
t------1 mmmm + 3

The illustration shows execution of LD A.I:

LD A.I
~

ED 57
Move the contents of the Interrupt Vector register to the Accumulator. and reflect Interrupt enable status in Panty/Overflow flag.
Suppose the Interrupt Vector register contains 7F16. and Interrupts are disabled. After
execu tlon of
LD A,I
Register A will contain 7F16. and P/O will be O.
LD A.R
~

ED 5F
Move the contents of the Refresh register to the Accumulator. The value of the Interrupt
flip-flop will appear in the Parity/Overflow flag.

3-94

LD A,(addr) S

F(

LOAD ACCUMULATOR FROM MEMORY USING

DIRECT ADDRESSING

Z AC Pia N C

I I I

Data
Memory

I I

YV

YV

ppqq

S,C

D,E
H,L

SP
PC

mmmm

IX

--

~:::r""mmm

IY

Program
Memory

I
I

I
R

+3

3A
qq

mmmm
mmmm+ 1

'I-"""::P;'::P--4

mmmm + 2
1 - - - - 1 mmmm + 3

LD A, (addrl
-.---.-3A

ppqq

Load the contents of the memorv bvte (addressed directlv bv the second and third
bvtes of the LD A, (addrllnstructlon object code) into the Accumulator. Suppose memorv byte 084A16 contains 2016, After the instruction
label

EQU

LD

A,(Iabef)

084AH

has executed, the Accumulator will contain 2016Remember that EQU is an assembler directive rather than an instruction: It tells the Assembler to use the 16-bit value 084A16 wherever the label appears,
The instruction
LD

A,(Iabef)

IS equivalent to the two instructions

LD
LD

HUabel
A(HL)

When vou are loading a single value from memory, the LD A'(Iabef) instruction IS preferred: It uses one instruction and three object program bytes to do what the LD HUabel.
LD A,(HL) combination does In two instructions and four object program bytes, Also,
the LD HUabeL LD A,(HL) combination uses the Hand L registers, which LD A,(Iabef)
does not.

3-95

LD A,(rp) - LOAD ACCUMULATOR FROM MEMORY LOCATION

ADDRESSED BY REGISTER PAIR
S

FI

Z ACP/0N C

I I

I !

A
S.C

yy

D.E

}-.sc or DE contain ppqq

H.l

SP
PC

mmmm

-.

..(;,mmm+V

IX

Data
Memorv

yy

ppqq

_t

Program

Memorv

IY
I
R

OOOxl0l0 mrnmm
mmmm+l
mmmm+2
1 - - -...... mmmm + 3

LDA,(rp)

IIw
........

o if register

palr=BC
1 if register palr=DE

Load the contents of the memory byte (addressed by the BC or DE register palrllnto the
Accumulator.
Suppose the B register contains OB16, the C register contains 4A16' and memory byte
084A16 contains 3A16. After the instruction
LD A,(BC)
has executed, the Accumulator will contain 3A16.
Normally, the LD A,(rp) and LD rp,data will be used together. since the LD rp,data instruction loads a 16-bit address Into the BC or DE registers as follows:
LD
LD

BC,084AH
A.(BC)

3-96

LD dst,src S

FI

MOVE CONTENTS OF SOURCE REGISTER TO

DESTINATION REGISTER

Z AC Pia N

Data

I I I I I. I

B.C
D.E
H.L

B. C.

Hir L

D. E. H. L

mmmm

IX
IY

D. E.

A.

~""~'A'C

SP
PC

Register

:;;;O:-::J.~mmm + ~

~
Program

Memorv

01dddsss

I
I

mmmm

I--_ _--tmmmm + 1
~---fmmmm + 2
~---fmmmm + 3

ill
01 ddd sss
~

000
001
010
011
100
101
111

for
for
for
for
for
for
for

dst
dst
dst
dst
dst
dst
dst

or
or
or
or
or
or
or

src=B
src=C
src=D
src=E
src=H
src=L
src=A

The contents of any designated register are loaded into any other register.
For example:
LD A.B
loads the contents of Register B into Register A.

LD L.D
loads the contents of Register 0 into Register L.
LD C.C
does nothing. since the C register has been specified as both the source and the
destination.

3-97

LD
LD
LD
LD

HL,(addr) -LOAD REGISTER PAIR OR INDEX REGISTER

rp, (addr)
FROM MEMORY USING DIRECT ADDRESSING
IX,(addr)
IY,(addr)
S

Z AC Pia N C

Data

I ! I ! ! ! I

B.C
D.E
H.L
SP
PC

Memorv

xx

ppqt

xx

yy

, mmmm+3

I-..~::I..

mmmm

IX
IY
I

ppqq

yy

Program
Memorv

2.A
qq
pp

mmm m
mmm m+ 1
mmm m+2.
mmm m+3

The illustration shows execution of LD HL(ppqq):

LD HL.addr
-----...-2A ppqq
Load the HL register pair from directly addressed memory location.
Suppose memory location 400416 contains AD16 and memOry location 400516 contains 1216. After the instruction
LD HL.(4004H)
has executed. the HL register pair will contain 12AD16.

~
ED 01 dd 1011 ppqq
00
01
10
11

for
for
for
for

rp is register pair Be
rp is register pair DE
rp IS register pair HL
rp IS Stack Pointer

Load register pair from directly addressed memory.

Suppose memory location 49FF16 contains BE16 and memory location 4A0016 contains 3316. After the Instruction
LD DE.(49FFH)
has executed. the DE register pair will contain 33BE16.

------LD IX.(addr)

DO 2A ppqq

Load IX register from directly addressed memory.

3-98

Suppose memory location 011116 contains FF16 and memory location 011216 contains 5616. After the instruction
LO IX.(0111 H)
has executed. the IX register will contain 56FF16
LO IY.(addrl

--.,---.,FD 2A ppqq
Load IY register from directly addressed memory.
Affects IY register instead of IX. Otherwise Identical to LO IX(addr).

lD I,A lD R,A
S
F

lOAD INTERRUPT VECTOR OR REFRESH

REGISTER FROM ACCUMULATOR

Z AC Pia N

Data

I I I I I I I
xx

B.C
D.E
H.L

SP
PC

mmmm

IX
IY
IV

~~mmm+!)

Program

Memory

ED
4F

mmmm
mmmm+!

.....- - - , 1 mmmm + 2
.....- - - , 1 mmmm + 3

The illustration shows execution of LO R.A:

LO R.A
~

ED 4F

Load Refresh register from Accumulator.

Suppose the Accumulator contains 7F16. After the Instruction
LO R.A
has executed. the Refresh register will contain 7F 16
LO I.A
~

ED 47
Load Interrupt Vector register from Accumulator.

3-99

LD reg,data S

LOAD IMMEDIATE INTO REGISTER

Z AC P/O N C

Data

I I I I I I I

}-""'''"''''" ,.

A
B.C
D.E
H.L

SP
PC

Register A. B. C.
D. E, H or L

mmmm

IX
IY
I
R

.- :l~mmm+~

Program

Memory
OOxxxll0
YY

mmmm
mmmm+ 1
mmmm+2

Il-----;lmmmm + 3

00 xxx 110 VV
000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Load the contents of the second object code byte into one of the registers.
When the instruction
LD A,2AH
has executed, 2A16 is loaded into the Accumulator.

3-100

lD rp,data-lOAD 16 BITS OF DATA IMMEDIATE INTO

lD IX,data
REGISTER
lD IV,data
S

Z ACP/ON

r:

H I I I I I I
A

B.C
D.E
H.L

SP
IX
IY

HL or

SO..,
'"ppqq
" into
SP. Load
sele::::estina tion

I
I

~
emorv

--~:I~mmm+v

mmmm

PC

Dala

Program
Memorv

OOXX0001
qq
pp

mmmm
mmmm+l
mmmm+2
mmmm+3

The illustration shows execution of LD rp,data:

~
0 0 xx 0001 ppqq
-,.-

00
01
10
11

for
for
for
for

rp IS register pair Be
rp is register pair DE
rp is register pair HL
rp is Stack POinter

Load the contents of the second and third object code bytes Into the selected register
pair. After the Instruction
LD SP,217AH
has executed, the Stack POinter will contain 217A 16
LD IX. data

'-v-" -,.-

DD 21 ppqq
Load the contents of the second and third object code bytes into the Index register IX.
LD IY.
--.-

data

-,.-

FD 21 ppqq
Load the contents of the second and third object code bytes Into the Index Register IY
Notice that the LD rp,data Instruction

IS

equivalent to two LD reg. data instructions.

For example:
LD

HL.032AH

is equivalent to
LD
LD

H,03H
L,2AH

3-101

LD reg,(HL) - LOAD REGISTER FROM MEMORY

LD reg, (IX +disp)
LD reg, (lY +disp)
S

Z AC Pia N C

Data

FI I I I I 0
A

}--''';'W
A. ,. C.
D. E. H or L

8.C

D.E

.......

H.L

SP
PC

mmmm

IX

ppqq

..r

---:I mmmm + 3
I-

~
yy

+d
ppqt

Program

Memory

IY

I
R

I
I

~pqq
-

d)::;:

DD
mmm m
01xxxll0 mmm m+ 1
mmm m+2
d
mmm m+3

The illustration shows execution of LD reg.(IX+disp):

~~.
DDOlxxxll0d

-.-

000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Load specified register from memory location (specified by the sum of the contents of
the IX register and the displacement digit tiL
Suppose ppqq=400416 and memory location 401016 contains FF16. After the instruction
LD B(lX+OCH)
has executed. Register B will contain FF16.

~J
FD 01 xxx 110 d

T...

. .,...... same as for LD reg.(IX+disp)

ThiS instruction IS Identical to LD reg.(IX+displ. except that It uses the IY register instead of the IX register.

3-102

m
0 1xxxll0

-.-

' - l - - - - I.... same as for LD reg,(lX+disp)

Load specified register from memory location (specified by the contents of the HL
register paid.

lD SP,Hl- MOVE CONTENTS OF Hl OR INDEX REGISTER

lD SP,IX
TO STACK POINTER
lD SP,IY
S

Z AC P/O N C

Data

FI I I I I I I
A
S.C

D.E
H.L

SP
PC

qq

pp

l:;)"
_ --,. mmmm + 1

mmmm

IY
I

Program
emorv

IX

I
I

~
F9

The illustration shows execution of LD SP,HL.

LD SP,HL
'-v-"

F9
Load contents of HL Into Stack POinter.
Suppose pp=0816 and qq=3F16. After the Instruction
LD SP,HL
has executed, the Stack POinter will contain 083F16
LD SP.IX
'-v-"

DD F9
Load contents of Index Register IX into Stack POinter.
LD SP,IY
~

FD F9
Load contents of Index Register IY Into Stack Pointer.

3-103

mmmm
mmmm+ 1
mmmm+2
mmmm+3

LD (addr),A S

STORE ACCUMULATOR IN MEMORY USING

DIRECT ADDRESSING

Z AC Pia N C

Data
Memorv

o::or::n

yy

yy

ppqq

B,C
D,E
H,L
SP

...r

I
I

Program
Memorv

..... -

IX
IY
I

~mmmm+3

mmmm

PC

32
qq
pp
1----;

mmmm
mmmm+1
mmmm+2
mmmm + 3

kr
32 ppqq

Store the Accumulator contents In the memory byte addressed directly by the second
and third bytes of the LD (addrl,A Instruction object code,
Suppose the Accumulator contains 3A16. After the instruction
label

EQU

084AH

LD

Oabell.A

has executed, memory byte 084A16 will contain 3A16'

Remember that EQU IS an assembler directive rather than an InstructioTi. It tells the Assembler to use the 16-bit value 084AH whenever the word "label" appears.
The instruction
LD (addrl,A
IS equivalent to the two instructions
LD H,label
LD (HL),A
When you are storing a Single data value In memory, the LD Oabell.A instruction IS
preferred because It uses one instruction and three object program bytes to do what the
LD H(label), LD (HL).A combination does In two Instructions and four oblect program
bytes. Also, the LD H(label), LD (HL).A combination uses the Hand L registers, while the
LD (labell.A Instruction does not.

3-104

lD (addr),Hl-STORE REGISTER PAIR OR INDEX

lD (addrl.rp
REGISTER IN MEMORY USING DIRECT
lD (addr) ,xv
ADDRESSING

Z AC Pia N C

I I I 0

Data
Memorv

yy

ppqq

B,C

xx

~ ppqq

D.E
H.L

SP
PC

/'

yy

xx

.-

mmmm

Program
Memory

t-.--:I mmmm + 4

......

IX
IY
I

ED
01010011
qq
pp

+1

rnmmm

mmmm+1
mmmm+2
mmmm+3

The illustration shows execution of LD (ppqql.DE:

lZED 01 xx 0011 ppqq

.......
00
01
10
11

for
for
for
for

rp
rp
rp
rp

is register pair BC
is register pair DE
IS register pair HL
is Stack Pointer

Store the contents of the specified register pair in memory. The third and fourth object
code bytes give the address of the memory location where the low-order byte is to be
written. The high-order byte is written into the next sequential memory location.
Suppose the BC register pair contains 3C2A 16. After the Instruction
label

LD

EQU

084AH

(Iabell.BC

has executed. memory byte 084A16 will contain 2A16. Memory byte 084B16 will contain 3C16.
Remember that EQU is an assembler directive rather than an instruction: it tells the Assembler to use the 16-bit value 084A16 whenever the word "label" appears.

~
22 ppqq

This is a three-byte version of LD (addrl,rp which directly specifies HL as the source

register pair.

3-105

~
DO 22 ppqq

Store the contents of Index register IX In memory. The third and fourth object code
bytes give the address of the memory location where the low-order bvte is to be Written. The high-order byte is written into the next sequential memory location.

rr
FD 22 ppqq

This instruction is identical to the LD (addrl.lX instruction. except that it uses the IY
register instead of the IX register.

3-106

LD (HL),data - LOAD IMMEDIATE INTO MEMORY

LD (lX+disp),data
LD (lY +disp) ,data
S
F

Z ACP/ON C

..

D:IIDJ

S.C

D.E

,. -

H.L

SP

~mmmm+4

mmmrn
ppqq

PC

IX

~pqq+drl......

IV

I
R

'--

Data
Memory

xx

+d

~l

Program

Memory

DO
36
d

xx

mmm m
mmm m+ 1
mmm m+2
mmm m+3

The illustration shows execution of LD (lX+d),xx:

--

LD (IX +disp),data

--..-

-.-

DO 36

xx

Load Immediate Into the Memory location designated by base relative addressing.
Suppose ppqq=540016. After the instruction
LD (lX+91.FAH
has executed. memory location 540916 will contain FA16.

-.---

LD (lY+disp),data

--..FD 36
This instruction
IX register.

IS

xx

identical to LD (IX +disp),data. but uses the IY register instead of the

LD (HU.data
---...-'

-.-

36

xx

Load Immediate Into the Memory location (specified by the contents of the HL register
pair),
The Load Immediate Into Memory instructions are used much less than the Load Immediate Into Register instructions.

3-107

LD (HLl.reg - LOAD MEMORY FROM REGISTER

LD (IX+disp) ,reg
LD (lY+disp),reg
S ZACP/ON

Data

FcrJ I I n
Contents of A, B,
C. 0, E, H or L
>----is Vy

B.C
D.E
H,L

qq

pp

SP
PC
IX

YV

mmmm

~mmmm+~

Program
Memorv

IY

01110xxx

mm mm
mm mm+l
mmmm+2

1----; mmmm + 3
The illustration shows execution of LD (HU.reg:

T1
01110xxx

000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg =H
reg=L
reg=A

Load memory location (specified by the contents of the HL register pair) from specified
register.
Suppose ppqq=450016 and Register C contains F916. After the instruction
LD (HU,C
has executed, memory location 450016 will contain F916.

J:Y
DD 01110 xxx

... same as for LD (HU,reg

Load memory location (specified by the sum of the contents of the IX register and the

3-108

displacement value dl from specified register.

LD (IY+disp).reg

:LI
FD 0111 0 xxx

T'----_..... same as for LD (HU.reg

This instruction IS identical to LD (IX+disp).reg. except that It uses the IY register instead of the IX register.

LD (rp),A -LOAD ACCUMULATOR INTO THE MEMORY

LOCATION ADDRESSED BY REGISTER PAIR
S

Z ACP/Q N C

I I I I I I I

Data
Memory

yy

B.C

D.E
H.L

contain ppqq

--.,- 0

SP
PC
IX

....BC mOE

mmmm

~-:J..mmmm+

yy

Program
Memorv

ppqq

IV

I
R

OOOXOO10

mmmm
mmmm+ 1
mmmm+2

1----1 mmmm + 3
LD (rp).A

il1

o if register palr=BC
1 if register pair=DE

Store the Accumulator in the memory byte addressed by the BC or DE register pair.
Suppose the BC register pair contains 084A16 and the Accumulator contains 3A16.
After the instruction
LD (BC).A
has executed. memory byte 084A 16 will contain 3A 16.
The LD (rp).A and LD rp.data will normally be used together. since the LD rp.data instruction loads a 16-blt address into the BC or DE registers as follows:
LD BC.084AH
LD (BC).A

3-109

LDD -

TRANSFER DATA BETWEEN MEMORY LOCATIONS,

DECREMENT DESTINATION AND SOURCE ADDRESSES

Set if BC-l '" O. reset otherwise

S

Z ACP/O N C

Data

FO 101 IIJ

ppqq-l
ppqq

B.C----t-t--.....----uu---"f"':jl{'
rrss-l

D.E 1I-__~r::::r---4---~ss~--_t:::_:_-~~--1

rrss

.LII-_ _.!:.PP!::..-_ _L _ _.:1:qq::!...._ _"~Iy--~

SPII~==_----PC Jt....;.m;;;m.;;;m~m~
_t....

...

IX

IY

--1

11.,..

-;

ED

II---;:A~8-'1I

t---,
1---,

mmmm
mmmm + 1
mmmm + 2
mmmm + 3

LDD

--.,-ED A8
Transfer a byte of data from memory location addressed by the HL register pair to
memory location addressed by the DE register pair. Decrement contents of register
pairs BC., DE. and HL.
Suppose register pair BC contains 004F16. DE contains 454516. HL contains 201216.
and memory location 201216 contains 1816. After the instruction
LDD
has executed. memory location 454516 will contain 1816. register pair BC will contain
004E16. DE will contain 454416. and HL will contain 201116.

3-110

LDDR-TRANSFER DATA BETWEEN MEMORY

LOCATIONS UNTIL BYTE COUNTER IS
ZERO. DECREMENT DESTINATION AND
SOURCE ADDRESSES
LDDR
"-v-'

ED B8

This instruction IS Identical to LDD, except that It IS repeated until the BC register pair
contains zero, After each data transfer, Interrupts will be recognized and two refresh Cycles will be executed,
Suppose we have the following contents In memory and register pairs:
Register/Contents
HL
DE
BC

Location/Contents

201216
454516
000316

201216 1816
201116 AA16
2010162516

After execution of
LDDR
register pairs and memory locations will have the following contents:
Register/Contents
HL
DE
BC

200916
454216
000016

Location/Contents

Location/Contents

201216
201116
201016

454516
454416
454316

1816
AA16
2516

1816
AA16
2516

This instruction is extremely useful for transferring blocks of data from one area of
memory to another,

3-111

LDI- TRANSFER DATA BETWEEN MEMORY

LOCATIONS. INCREMENT DESTINATION AND
SOURCE ADDRESSES
Set if BC-1 ,p O. reset otherwise

S Z ACP'ON C

FI

101

liD

Ar----:'~-~t--~~----ir:;
......,f-__..;:u:;;,u_ _--4~/'"

B.C II-_ _....:t::.,t_ _

1.==='1

D.E ~_ _~r~r_ _-l~--~s;s---I~---=~=---'"'I

.L~-....!Pp~-.J.--~q~q
_ _~~.........--~
SP
PC I m...m
...m
......
m
--t~

II------- --------f..
---1

IX 11-

IY ....

..,..

--I

ED
mmmm
t--;A:;:.O-.... mmmm + 1

1-----4 mmmm + 2
t - - -.... mmmm + 3

LDI

-...ED AO
Transfer a byte of data from memory location addressed by the HL register pair to
memory location addressed by the DE register pair. Increment contents of register pairs
HL and DE. Decrement contents of the BC register pair.
Suppose register pair BC contains 004F16. DE contains 454516. HL contains 201216.
and memory location 201216 contains 1816. After the instruction
LDI
has executed. memory location 454516 will contain 1816. register pair BC will contain
004E16. DE will contain 454616. and HL will contain 201316.

3-112

LDIR -

TRANSFER DATA BETWEEN MEMORY

LOCATIONS UNTIL BYTE COUNTER IS
ZERO. INCREMENT DESTINATION AND
SOURCE ADDRESSES
LOIR
'-v-'

ED BO
This Instruction is Identical to LOL except that It IS repeated until the BC register pair
contains zero. After each data transfer. Interrupts will be recognized and two refresh cycles will be executed.
Suppose we have the following contents In memory and register pairs:
Register/Contents
HL
DE
BC

Location/Contents

201216
454516
000316

2012161816
201316 C016
201416 F016

After execution of
LOIR
register pairs and memory will have the following contents:
Register/Contents
HL
DE
BC

Location/Contents

201516
454816
000016

201216
201316
201416

Location/Contents

1816
C016
F016

454516
454616
454716

1816
C016
F016

This instruction is extremely useful for transferring blocks of data from one area of
memory to another.

NEG S

NEGATE CONTENTS OF ACCUMULATOR

c

ZACP/ON

F~IJ
A

- , Xx'+
1 )

=--J.

xx

B.C
D.E
H.L

SP
PC
IX
IY
I
R

Data

mmmm

~mmmm+2

........

Program
Memory

.ED
44

mmmm
mmmm+ 1
I----~ mmmm + 2
I--_ _~ mmmm + 3

Negate contents of Accumulator. This is the same as subtracting contents of the Accumulator from zero. The result is the two's complement. 80H will be left unchanged.
Suppose xx=5A 16. After the instruction

NEG
has executed. the Accumulator will contain A616
5A
0101
Two's complement = 1 01 0

1010
01 1 0

3-113

NOP 5

NO OPERATION
Z AC Pia N C

Data

Fcr::IIIIJ
A

B.C

D.E
H.L

SP
PC

mmmm

-"

IX
IY

~
Program

~
emorv

00

~~~mmm+l

NOP
--.--

mmmm
mmmm+l
mmmm+2

mmmm+3

00
This IS a one-byte instruction which performs no operation, except that the Program
Counter IS Incremented and memory refresh continues. This instruction is present for
several reasons:
1)

A program error that fetches an object code from non-existent memory will fetch
00. It is a good idea to ensure that the most common program error will do nothing.

2)

The NOP Instruction allows you to give a label to an object program byte:
HERE NOP

3)

To fine-tune delay times. Each NOP instruction adds four clock cycles to a delay.

NOP is not a very useful or frequently used instruction.

3-114

OR data S

OR IMMEDIATE WITH ACCUMULATOR

ZACP!ON C

Data

F~
A

:::;;o::-=C.xx OR YY~

xx

B,C
D,E

H,L

SP
PC

mmmm

......

IX
IY
I

mmmm+2

F6
yy

Program
Memory

OR

data

F6

yy

mmmm
mmmm+ 1
mmmm+2
mmmm+3

OR the Accumulator with the contents of the second instruction object code byte.
Suppose xx=3A 16 After the instruction
OR 7CH
has executed. the Accumulator will contain 7E16'
3A
7C

sets S to

0011
01 1 1

1010
1 100

0111

1110

o.....J

LSiX 1 bits, set PIO to 1

LNon-zero result. set Z to 0

This IS a routine logical instruction: it

instruction

IS

often used to turn bits" on" For example, the

OR SOH
will unconditionally set the high-order Accumulator bit to 1.

3-115

OR reg S

OR REGISTER WITH ACCUMULATOR

ZACP/DN

xx

.......xxOR Y~
.

ff

'XIXI11xloioi

} ,"",.1.

A B.
-IIIo-C, 0, E, H or L
IS yy

B,C
D,E
H,L

SP
PC
IX

Data

'-

".

~mmmm+\

mmmm

IY

I
R

OR
-..--

Program
Memory

10110xxx mmmm
1l-_ _--4mmmm + 1
11-_ _-1 mmmm + 2
11-_ _-1 mmmm + 3

reg

10110 xxx

0i50 for reg=B

001
010
011
100
101
111

for
for
for
for
for
for

reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Logically OR the contents of the Accumulator with the contents of Register A. B. C. D.

E. H or L. Store the result in the Accumulator.
Suppose xx=E316 and Register E contains A816' After the instruction
OR E
has executed. the Accumulator will contain EB16.
E3
A8 =

1 1 10
1010
1 110

001 1
1000
101 1

1 sets S to 1.-J

LSix 1 bits. set P!O to 1

LNon-zero result. set Z to 0

3-116

OR (HL) -OR MEMORY WITH ACCUMULATOR

OR (IX+disp)
OR (lY+disp)
S ZACP/ON
F

Data
Memory

IxlxllIxlolo'
xx

1.....oL.....

B.C
D.E
H.L
SP
PC

:I~XORyy~

P qq

qq

PP

:::::;; -.- mmmm + 1

mmmm

.........

IX
IY
I
R

yy

Program
Memory

~
B6

The illustration shows execution of OR (HL):

OR (HL)

mmm

m
m mmm+l
mmmm+2
mmmm+3

86
OR contents of memory location (specified by the contents of the HL register pair) with
the Accumulator.
Suppose xx=E316. ppqq=400016. and memory locatIOn 400016 contains AS16. After
the instruction
OR

(HU

has executed. the Accumulator will contain E816.

E3
AS =

1 1 10
001 1
_1.:..0..;.1..;.0_..;.1",,-0.:..0..:.,0
1 1 10
101 1
LSiX 1 bits. set PIO to 1

1 sets S to 1......J

---

LNon-zero result. set Z to

OR (lX+disp)

DD 86

-,-

OR contents of memory location (specified by the sum of the contents of the IX register
and the displacement value d) with the Accumulator.
OR (lY+disp)
-.,~

FD 86

This instructIOn IS Identical to OR (lX+displ. except that It uses the IY register instead of
the IX register.

3-117

OUT (e) ,reg -

OUTPUT FROM REGISTER

Z AC PIO N C

I I I

IJ 0
A

I
I
yy

B.C
D.E

t
}__,,,,,J~ ",
I/O port yy I

D. E, H or L

H.L
SP

mmmm

PC
IX
IY
I

~mmm+.D

~
Program

Memorv

ED

I
I

Data

01xxx001

r------1

l--i
ED 01 xxx 001
000
001
010
all
100
101
111

for reg=B
for reg=C
for reg=D
for reg=E
for reg=H
forreg=L
for reg=A

Suppose yy= 1F16 and the contents of Hare AA16. After the execution of
OUT (Cl.H
AA16 will be

In

the buffer of I/O port 1F16.

3-118

mmmm
mmmm + ,
mmmm+2
mmmm + 3

OUTD -

s
F

OUTPUT FROM MEMORY. DECREMENT ADDRESS

Z AC

Pia N C

ITIEEI u I 11 ]

xx-l

Lf<!0IW

Ij

B.C
D.E

xx

H.L

PP

SP
PC

yy

--

--

~mmm+2

IX

I
I

I
ppq q

~ Ppqq-l)

qq

mmmm

IY
I

Data
Memorv

Program
Memorv

ED
AB

1----1

mmmm
mmmm+l
mmmm+2
mmmm + 3

aUTO
-.,-..'

ED AB
Output from memory location specified by HL to I/O port addressed by Register C.
Registers Band HL are decremented.
Suppose xx=OA16' yy=FF16. ppqq=500016. and memory location 500016 contains
7716. After the instruction
aUTO
has executed. 7716 will be held In the buffer of I/O port FF16. The B register will contain 0916, and the HL register pair 4FFF16.

OTDR -

OUTPUT FROM MEMORY. DECREMENT ADDRESS,

CONTINUE UNTIL REGISTER 8=0
OTOR
-.,-..'

ED BB
OTOR is identical to aUTO, but is repeated until Register B contains 0.
Suppose Register B contains 0316, Register C contains FF16' and HL contains 500016.
Memory locations 4FFE16 through 500016 contain:
Location/Contents
4FFE16
4FFF16
500016

CA16
1B16
F116

After execution of
OTOR
register pair HL will contain 4FF016. Register B will contain zero, and the sequence
Fl16' 1B16, CA 16 will have been written to I/O port FF16.
ThiS instruction IS very useful for transferring blocks of data from memory to output
devices.

3-119

OUTI- OUTPUT FROM MEMORY. INCREMENT ADDRESS

S

Z ACP/O N C

xx-1

+
+

lulxlulul1l I
A
S,C

I/O port vv I

yy

xx

D.E
H,L

Data
Memorv
pp qq

..-r..... -

ppqq+l

pp

qq

SP
PC

mmmm

io!J!Y

--....

Program

""'~mmm+2

IX

IV
I

Memorv

~
ED

A3

OUTI
-.,-'

mmm

m
mmmm+l
mmmm+2
mmmm+3

ED A3
Output from memory location specified by HL to I/O port addressed by Register C.
Register 8 IS decremented and the HL register pair IS Incremented.
Suppose xx=OA16, yy=FF16' ppqq=500016, and memory location 500016 contains
7716. After the instruction
OUTI
has executed, 7716 will be held in the buffer of I/O port FF16. The 8 register will contain 0916 and the HL register pair will contain 500116,

OTIR -

OUTPUT FROM MEMORY. INCREMENT ADDRESS,

CONTINUE UNTIL REGISTER 8=0
OTIR

-.,ED 83
OTIR is Identical to OUTI. except that It IS repeated until Register 8 contains O.
Suppose Register 8 contains 0416, Register C contains FF16' and HL contains 500016.
Memory locations 500016 through 500316 contain:
Location/Contents
5000 16 CA16
500116 18 16
5002 16 81 16
5003 16 AD16
After execution of
OTIR
register pair HL will contain 500416, Register 8 will contain zero and the sequence
CA16. 1816, 8116 and AD16 will have been written to I/O port FF16.
This instruction IS very useful for transferring blocks of data from memory to an output
device,

3-120

OUT (port),A -

s
F

OUTPUT FROM ACCUMULATOR

Z AC P/O N C

0 .. 1 I CO

A
S.C
D.E

~mmm+!)

H.L

SP
PC

mmmm

IX
IY
I

I/O port yy ~

Data

~
Program
Memory
D3
yy

OUT

mmmm
mmmm+l
mmmm+2
mmmm+3

(portl.A

1/1
03

yy

Output the contents of the Accumulator to the I/O port Identified by the second OUT instruction oblect code byte.
Suppose 3616 IS held in the Accumulator. After the instruction
OUT (lAHl.A
has executed. 3616 will be in the buffer of I/O port 1A16.
The OUT instruction does not affect any statuses. Use of the OUT Instruction is very
hardware-dependent. Valid 1/0 port addresses are determined by the way In which 1/0
logiC has been Implemented. It is also possible to deSign a microcomputer system that
accesses external logiC uSing memory reference instructions with specific memory addresses. OUT instructions are frequently used In special ways to control microcomputer
logic external to the CPU.

3-121

POP rp -

READ FROM THE TOP OF THE STACK

POP IX
POPIY
S

Z ACP/ON

Data

I I I I I ! I

Memory

qq
S.C

D.E
H.L

SP

SSSS

PC
IX
IV
I

mmmm

pp

~A

5555
5555 + 1
5555 + 2

5555+2)

~-=

Program

"'-~:J. mmmm + 1

Memory

11000001

mmmm

mmmm+1
mmmm+2
11----1 mmmm + 3

The illustration shows execution of POP BC.

POP rp

1Z11 xx 0001
00
01
10
11

for
for
for
for

rp
rp
rp
rp

s
s
s
s

reg
reg
reg
reg

ster
ster
ster
ster

pa
pa
pa
pa

r
r
r
r

BC
DE
HL
A and F

POP the two top stack bytes in10 the designated register pair.
Suppose qq=0116 and pp=2A16 Execution of
POP HL
loads 0116 into the L register and 2A 16 Into the H register. Execution of the instruction

paPAF
loads 01 Into the status flags and 2A 16 Into the Accumu lator. Thus. the Carry status
will be set to 1 and other statuses will be cleared.
POP IX
~

DD El
POP the two top stack bytes Into the IX register.
POP IY
~

FD E1
POP the two top stack bytes Into the IY register.
The POP Instruction IS most frequently used to restore register and status contents
which have been saved on the stack; for example. while servicing an Interrupt.

3-122

PUSH rp-WRITE TO THE TOP OF THE STACK

PUSH IX
PUSHIY

I I I
A
B,C
D,E
H,L
SP

SSSS

PC

mmmm

IX
IV

I
R

Data
Memory

CTJ

qq
pp

ssss-2
ssss-l
ssss

SSSS-2)

- ')

//

~-,.
_,......; mmmm+2

Program

ppqq

Memory

FD

E5

rnmmm
mmmm+ 1

1----1 mmmm + 2
11----1 mmmm + 3

The illustration shows execution of PUSH IY:

PUSH IY
'-v-'
FD E5
PUSH the contents of the IY register onto the top of the stack,
Suppose the IY register contains 45FF16' Execution of the instructIOn
PUSH IY
loads 4516, then FF16 onto the top of the stack,
PUSH IX
'-v-'

DD E5
PUSH the contents of the IX register onto the top of the stack,

JK
11

xx

00
01
10
11

0101

for
for
for
for

rp
rp
rp
rp

is
IS
is
IS

register
register
register
register

pair Be
pair DE
pair HL
pair A and F

PUSH contents of deSignated register pair onto the top of the stack,
Execution of the instruction
PUSH AF
loads the Accumulator and then the status flags onto the top of the stack,
The PUSH Instruction is most frequently used to save register and status contents: for
example, before servicing an Interrupt

3-123

RES b,reg -

RESET INDICATED REGISTER BIT

Data

S Z AC PIa N C

FD_J I I I ,
A

YYYYYYYY

B.C
D.E

&

H.L

SP
PC

~mmm+:D

mmmm

IX
IY

I
R

Program
Memory
CB

I
I

10bbbxxx

mmmm
mmmm + 1

mmmm+2

1----1 mmmm + 3

-ll\
CB

10 bbb xxx

Bit

-.- -.bbb ~

1
2
3
4
5
6

000
001
010
011
100
101
110

111

000
001
010
011
100
101
111

Register

B
C

o
E
H
L

Reset Indicated bit within specified register.

After the instruction
RES 6.H

has executed. bit 6 In Register H will be reset. (Bit 0

3-124

IS

the least significant bit.)

RES b,(HL) - RESET BIT b OF INDICATED MEMORY POSITION

RES b,(lX+disp)
RES b, (lY+disp)
S ZACP/ON

F'

Data

I I I I I I
0

B.C
D.E

H.L
SP

PC
IX
IY
I
R

mmmm

~mmmm+4

ppqq

CPpqq+d~

YvvvYYYV

ppqq+ d

Program
Memorv

mmm m
mmm m+l
mmm m+2
d
10bbbll0 mmm m+3

DO

CB

mmm m+4

The illustration shows execution of SET b.(IX+disp). Bit 0 IS execution of SET

b.(IX+disp). Bit 0 IS the least significant bit.

~
D DCB d 10 bbb 110
bbb
000
001
010
011
100
101
110
111

Bit Reset
0
1
2
3
4
5
6
7

Reset Indicated bit within memory location Indicated by the sum of Index Register IX
and d.
Suppose IX contains 411016. After the instruction
RES 0.(IX+7)
has executed. bit 0 In memory location 411716 will be O.

~
FDCB d 10 bbb 110

-.-

bbb IS the same as In RES b.(IX+disp)

This Instruction IS identical to RES b. (IX+displ. except that it uses the IY register instead

3-125

of the IX register.
RES b.(HU

III
CB 10 bbb 110

bbb IS the same as in RES b.(IX+disp)

Reset Indicated bit within memory location Indicated by HL.
Suppose HL contains 444416. After execution of
RES 7.(HU
bit 7 in memory location 444416 will be O.

RET 5

RETURN FROM SUBROUTINE

ZACP/ON

Fern

Data

I I

emorv

PP

B.C
D.E
H.L
SP
PC

xxxx

mmmm

IX
IY
I

r~
qq

- + 2)
_---i xxxx

~~).
RET

xxxx
xxxx+ 1
xxxx + 2

Program

emorv

~
C9.

mmmm
mmmm+l
mmmm+2
mmmm+3

C9
Move the contents of the top two stack bytes to the Program Counter: these two bytes
provide the address of the next Instruction to be executed. Previous Program Counter
contents are lost. Increment the Stack POinter by 2. to address the new top of stack.
Every subroutine must contain at least one Return (or conditional Return) instruction:
thiS IS the last instruction executed within the subroutine. and causes execution to
return to the calling program.

3-126

RET cond -

RETURN FROM SUBROUTINE IF CONDITION

IS SATISFIED

~
11 xxx 000

000
001
010
011
100
101
110
111

NZ
Z
NC
C
PO
PE
P
M

Condition
Non-Zero
Zero
Non-Carry
Carry
PantyOdd
Parity Even
Sign Positive
Sign Negative

Relevant Flag
Z
Z
C
C

Pia
PIO
S
S

This instruction IS Identical to the RET instruction. except that the return IS not executed unless the condition IS satisfied: otherwise. the instruction sequentially followIng the RET cond instruction will be executed.
Consider the instruction sequence:
CALL
AND

SUBR
7CH'-'
1
I
I

: :First subroutine instruction

I
: condition satisfied

R T

cond

----------'

condition not
satisfied

80H

After the RET cond is executed. if the condition IS satisfied then execution returns to the
AND instruction which follows the CALL. If the condition IS not satisfied. the OR instruction. being the next sequential instruction. IS executed.

3-127

RETI- RETURN FROM INTERRUPT

S Z ACP/0N

Data

I I I I I I I

Memorv

qq
pp

B.C
D.E
H.L
SP
PC

xxxx
mmmm

-...

IX
IV
I

XXXX
XXXX + 1
XXXX + 2

l:::XXX+2)

r
-

ppqq

I
RETI

Program
Memory

~
ED
4D

mmmm

mmmm+1
mmmm+2
mmmm+3

'-.,-'

ED 4D
Move the contents of the top two stack bytes to the Program Counter: these two bytes
provide the address of the next instruction to be executed. PrevIous Program Counter
contents are lost. Increment the Stack POinter by 2. and address the new top of stack.
This Instruction IS used at the end of an Interrupt service routine. and. in addition to
returning control to the Interrupted program. it is used to signal an I/O deVice that the
Interrupt (outine has been completed. The I/O deVice must provide the logic necessary
to sense the Instruction operation code: refer to An Introduction to Microcomputers: Volume 2 for a deSCription of how the RETI instruction operates With the Z80
familv of deVices.

3-128

RETN S

RETURN FROM NON-MASKABLE INTERRUPT

ZACP/ON

Data
Memory
qq
pp

FCIIII:D
A

e.C
D.E
H.L

SP
PC
IX
IY
I
R

xxxx
mmmm

.r

-.:..1

mmmm
mmmm+l
mmmm+2

xxxx +2

rL( - r
ppqq

Program
Memory
ED

mmmm

45

mmmm+ 1
mmmm+2
mmmm+3

--RETN

ED 45

Move the contents of the top two stack bytes to the Program Counter: these two bytes
provide the address of the next instruction to be executed. PrevIous Program Counter
contents are lost. Increment the Stack Pointer by 2 to address the new top of stack.
Restore the Interrupt enable logiC to the state It had prior to the occurrence of the nonmaskable Interrupt.
ThiS instruction IS used at the end of a service routine for a non-maskable interrupt. and
causes execution to return to the program that was interrupted.

3-129

RL reg -

ROTATE CONTENTS OF REGISTER LEFT

THROUGH CARRY

Z AC P/O N

Data

FIXIXIOIXIO,I
A
u.u

D.E

H.L

SP
PC

__ ~'::mmm0

mmmm

Program
Memorv

IX

IY
I

CB
00010001

mmmm
mmmm + 1
mmmm+2

11-----11 mmmm + 3

The illustratton shows execution of RL C.

-t\
CB 00010 xxx

000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Rotate contents of specified register left one bit through Carry.

Suppose D contains A916 and Carry=O. After the instruction
RL D
has executed. D will contain 5216 and Carry will be 1
Before

After

Register D

Carry

Register D

Carry

1101010011

[ill

ill

o sets S to 0......J
3 ones. set P/0 to 0

3-130

L.Non-zero result. set Z to 0

RL (HL) ROTATE CONTENTS OF MEMORY LOCATION

RL (lX+dispjLEFT THROUGH CARRY
RL (lY +disp)

A_----~f_

____I

11-------;-------;
11-------;-------;
H.L
SP 1 1 - - - - - - - - ' - - - - - - - ;

B.C
D.E

PC
IX
IY
I

II------m-m-m-=mr--------i-....r
t------j)pi)ipqi!cql------r--.........._ _""

------'"""'Ir--------;

aJ
Memory

DD
mmmm
CB
mmmm+ 1
d
mmmm+2
........_ _,..rr'lIII""il---;;,6;,.....--1 mmmm + 3

1-_ _-1 mmmm + 4

The illustration shows execution of RL (lX+disp):
RL (lX+disp)
--.--~

Rotate contents of memory location (specified by the sum of the contents of Index
Register IX and displacement Integer d) left one bit through Carry.
Suppose the IX register contains 400016. memory location 400716 contains 2F, 6. and
Carry is set to 1. After execution of the Instruction
RL (lX+7)
memory location 400716 will contain 5F16. and Carry is 0:
Before

After

Memory

Carry

Memory

Carry

1001011111

OJ

[QJ

0.J

o sets S to
6 ones. set P10 to 1

L.Non-zero result. set Z to 0

FD CB d 16

This Instruction
register.

IS

Identical to RL (IX+displ. but uses the IY register Instead of the IX

3-131

RL {HU
~

CB 16
Rotate contents of memory location (specified by the contents of the HL register pair)
left one bit through Carry.

RLA -

ROTATE ACCUMULATOR LEFT THROUGH CARRY

S ZACPION~
FI I 101 lo~

Data

B.C

D.E
H.L
SP
PC
IX
IY
I

,.

mmmm

~"""'"":l

mmmm+l

.......

I
I

~
Program

emorY

~
17

RLA

mmmm
mmmm+1
mmmm+2
mmmm+3

17
Rotate Accumulator contents left one bit through Carry status.
Suppose the Accumulator contains 2A16 and the Carry status is set to 1. After the instruction

RL.A
has executed. the Accumulator will contain F516 and the Carry status will be reset to 0:
Before

After

Accumulator
10 1 1 1 1 0 1 0

Carry

Accumulator

Carry

OJ

11111 01011

@]

3 -132

RLC reg S

ROTATE CONTENTS OF REGISTER LEFT CIRCULAR

Z AC P/O N

Data

FlxlxlOlxlOI ..........
A
B.C

H.L
SP
PC
IX
IY
I.
R

....... -::I.....::mmm +.:D

mmmm

Program

Memory
mmmm
mmmm + 1
mmmm+2
I - - - - i mmmm + 3
CB
00000011

The illustration shows execution of RLC E.

Jl
CB 000 00 xxx

000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Rotate contents of specified register left one bit. copying bit 7 Into Carry.
Suppose Register D contains A916 and Carry is 1. After execution of
RLC D
Register D will contain 5316 and Carry will be 1.
Before

After

Register D

Carry

11010 10011

IJJ

sets S to 0
4 ones. set P/0 to 1

3-133

Register D

Carry

ROTATE CONTENTS OF MEMORY LOCATION

RLC (HL)RLC (IX +disp) LEFT CIRCULAR
RLC (IV +disp)

Data

~]]TI[]rr2I:J.------------~L_J:i::Qiiiw::d._.J
Memory
A

B.C
D.E
H.L

:.~mmm+V
-

qq

pp

SP
PC

mmmm

IX

IY
I
R

ppqq

t
Program

Memory
mmm m
mmm m+1
mmmm+2
1------1 mmmm + 3

CB
06

The illustration shows execution of RLC (HL):

RLC (HL)
'-v-"

CB 06
Rotate contents of memory location (specified by the contents of the HL register paid
left one bit. copying bit 7 Into Carry.
Suppose register pair HL contains 54FF16 Memory location 54FF16 contains A516.
and Carry IS O. After execution of
RLC (HL)
memory location 54FF16 will contain 4B16. and Carry will be 1
Before

After

Memory

Carry

1101001011

@]

Memory

Carry

OJ

sets S to 0
4 ones, set P/0 to 1
RLC (IX+disp)
'-v-"-

Rotate memory location (specified by the sum of the contents of Index register IX and
displacement Integer d) left one bit. cOPYing bit 7 into Carry.
Suppose the IX register contains 400016. Carry is 1. and memory location 400716 contains 2F16. After the instruction
RLC (IX+7)

3-134

has executed. memory location 400716 will contain 5E16. and Carry will be 0:
Before

After

Memory

Carry

Memory

Carry

1001011111

III

@]

o.-J

sets S to
5 ones. set P/0 to 0

L.Non-zero result. set Z to 0

RLC (lY+disp)

'-.r-"-'-

This instruction
register.

RLCA S

FI

IS

identical to RLC (IX+disp). but uses the IY register Instead of the IX

ROTATE ACCUMULATOR LEFT CIRCULAR

Z AC P / ON

I 101 101

Data

B.C

D.E
H.L
SP
PC
IX
IY
I

--

__--:r mmmm + 1
.......

mmmm

Program ::
emory

~
07

I
I

RLCA
--..-

mmmm
mmmm+ 1
mmmm+2
mmmm+3

07

Rotate Accumulator contents left one bit. copying bit 7 Into Carry.
Suppose the Accumulator contains 7A16 and the Carry status
struction

IS

set to 1. After the in-

RLCA
has executed. the Accumulator will contain F416 and the Carry status will be reset to 0:
Before

After

Accumulator

Carry

Accumulator

Carry

1011110101

OJ

1111101001

@]

RLCA should be used as a logical instruction.

3-135

RLD S

ROTATE ONE BCD DIGIT LEFT BETWEEN

THE ACCUMULATOR AND MEMORY LOCATION
Z AC PIO N C

Data

F'xIXlolxlol I
x

B.C
D.E
H.L

SP
PC

PP

qq

_~~mmm+:D

mmmm

IX
IY
I

Program

Memory

ED
6F

PP1

.----1

mmm m
mmm m+ 1

mmmm + 2
t -_ _-lImmmm + 3

RLD
--.".-'

ED 6F
The four low-order bits of a memory locatIOn (specified by the contents of register pair
HL) are copied Into the four high-order bits of the same memory location. The prevIous
contents of the four high-order bits of that memory location are copied into the four
low-order bits of the Accumulator. The prevIous four low-order bits of the Accumulator
are copied Into the four low-order bits of the specified memory location.
Suppose the Accumulator contains 7F16. HL register pair contains 400016. and memory location 400016 contains 1216. After execution of the instruction

RLD
the Accumulator will contain 7116 and memory location 400016 will contain 2F16:
Before
Accumulator
7

After
Memorv

Accumulator

Memory

o::DJ~l ffi]
\.....F J\t_.,/ll~
...'

high-order

blt=O~::t-~~~-O

4 ones. set P/0 to 1

3-136

Non-zero result. set Z to 0

RR reg -

ROTATE CONTENTS OF REGISTER RIGHT THROUGH

CARRY
CIIII

'I

Z ACP/D N

Data

FlxlxlOlxlOI

D.E
H.L
SP
PC
IX
IY
I

-.

mmmm

::l. mmmm + 2

Program

Memory

CB

00011001

mmmm
mmmm + 1
mmmm+2

!----I,I mmmm + 3

The illustration shows execution of RR C:

L\
CB 00011 xxx

000
001
010
011
100
101
111

for reg=B
for reg=C
for reg=D
for reg=E
for reg=H
forreg=L
for reg=A

Rotate contents of specified register nght one bit through Carry.

Suppose Register H contains OF 16 and Carry

IS

set to 1. After the instruction

RR H
has executed. Register H will contain 8716. and Carry will be 1
Before
-

Register H

After

-Carry

Register H

Carry

100 00 111 11

1 sets S to 1
4 ones. set P/0 to 1

3-137

LNon-zero result. set Z to

RR (HL) -

ROTATE CONTENTS OF MEMORY LOCATION

RIGHT THROUGH CARRY
RR (IX+disp)
RR (lY +disp)
S Z ACPrON

Data

Memory

IXIXIOIXIOIVI

ppq q+d

B.C
D.E
H.L

SP
PC

~--.{"mmmm

mmmm

+l:)

IX

IY

Program
Memory

ppqq

FD
CB
d
lE

(Ppqq+d~

mmm m
mmm m+l
mmm m+2
mmm m+3
mmm m+4

The illustratIOn shows execution of RR (lY+disp):

RR (lY+~)

-.,,-'

J;;h

Rotate contents of memory location (specified by the sum of the contents of the IY
register and the displacement value dl right one bit through Carry.
Suppose the IY register contains 450016. memory locatIOn 450F16 containS 1D16. and
Carry IS set to O. After execution of the instruction
RR (lY+OFHl
memory location 450F16 will contain OE16. and Carry will be 1
Before

After

Memory

Carry

1000111011

@]

Memory

Carry

OJ

sets S to 0
3 ones. set P/0 to 0
RR (lX+disp)
-.,,-'

-.-

ThiS instruction is Identical to RR (lY+displ. but uses the IX register Instead of the IY
register.

3-138

RR (HU
~

CB lE
Rotate contents of memory location (specified by the contents of the HL register pair)
fight one bit through Carry.

RRA-ROTATE ACCUMULATOR RIGHT THROUGH CARRY

s
FI

Z AC P/O N

101

Data

101

B.C

D.E
H.L
SP
PC
IX
IY
I'

~...:mmm+)

mmmm

~
Program

~
emorY

1F

RRA

mmmm
mmmm+ 1
mmmm+2
mmmm+3

IF
Rotate Accumulator contents fight one bit through Carry status.
Suppose the Accumulator contains 7A16 and the Carry status IS set to 1. After the instruction
RRA
has executed. the Accumulator will contain B016 and the Carry status will be reset to

0:
After

Before
Accumulator
1011110101

Accumulator

Carry

11011 11011

@]

Carry

3-139

RRC reg -

ROTATE CONTENTS OF REGISTER RIGHT CIRCULAR

s Z AC PIO N C
FLXIXIOIXIOl

Data

A
B.C

D.E
SP
PC

~mmm+!)

mmmm

IX
IY
I

Program
Memory
CB
00001101

I
I

1----1

mmmm
mmmm + 1
mmmm+2
mmmm + 3

The illustration shows execution of RRC L:

RRC reg

l~
CB 00001 xxx

000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=O
reg=E
reg=H
reg=L
reg=A

Rotate contents of specified register right one bit clrcularry. copying bit 0 Into the Carry
status.
Suppose Register 0 contains A916 and Carry is O. After execution of
RRC 0
Register 0 will contain 0416. and Carry will be 1
Before

After

Register 0

Carry

Register 0

Carry

1101010011

[Q]

1110101001

IT1

, "" S "

,..re.,,,"

4 ones. set P/0 to 1

3-140

"'"It. '"

Z 100

RRC (HL)RRC OX +disp)

RRC OV +disp)
S

ZACP/ON

ROTATE CONTENTS OF MEMORV LOCATION

RIGHT CIRCULAR

Data
Memory

FXXOXO

p Pqq

B.C
D.E
H.L

qq

pp

--

SP
PC
IX

mmmm

IY

1\
R

+
:I~mmm+:D

Program
Memory

~
CB
OE

The illustration shows execution of RRC (HL):

mmm
mmm m+ 1
mmmm+2
mmmm+3

RRC (HU
~

CB OE
Rotate contents of memory location (specified by the contents of the HL register pair)
right one bit circularly. copYing bit 0 into the Carry status.
Suppose the HL register pair contains 450016. memory location 450016 contains
3416. and Carry is set to 1. After execution of
RRC (HL)
memory location 450016 will contain 1A16. and Carry will be 0:
Before

After

Memory

Carry

1001101001

OJ

Memory

o sets S to 0
3 ones. set PIO to 0

Carry

L.Non-zero result. set Z to 0

RRC (IX+disp)

'Xh

Rotate contents of memory location (specified by the sum of the contents of the IX

3-141

register and the displacement value d) right one bit circularly, copying bit 0 into the Carry status.
RRC (IY+disp)

This instruction IS identical to the RRC (IX+disp) instruction, but uses the IY register instead of the IX register.

RRCA -

s
FL I

ROTATE ACCUMULATOR RIGHT CIRCULAR

Z AC P/O N C

101

Data

101

B.C
D.E
H.L
SP
PC

-,

mmmm

~~~mmm+l

IX
IY

1
R

~
Program

emorY

~
OF

I
I

--RRCA

mmmm
mmmm+l
mmmm+2
mmmm+3

OF

Rotate Accumulator contents fight one bit circularly, cOPYing bit 0 into the Carry status.
Suppose the Accumulator contains 7A 16 and the Carry status is set to 1. After the instruction
RRCA
has executed, the Accumulator will contain 3016 and the Carry status will be reset to

0:
Before

After

Accumulator

Carry

Accumulator

Carry

1011110101

OJ

1001111011

[Q]

RRCA should be used as a logical instruction.

3-142

RRD S

ROTATE ONE BCD DIGIT RIGHT BETWEEN THE

ACCUMULATOR AND MEMORY LOCATION
AC p!a N C
Data

Ixlxlolxlol I

M~ry

r
I

I~~

B.C

D.E
H.L

SP
PC

qq

PP

"

Program
Memory

__~:.l mmmm + 2

mmmm

IX

P q

IY

I
R

ED
67

mm mm
mm mm+1
mmmm+2

.----I mmmm + 3

-RRD

ED 67

The four high-order bits of a memory location (specified by the contents of register pair
HL) are copied Into the four low-order bits of the same memory location. The previous
contents of the four low-order bits are copied into the four low-order bits of the Accumulator. The prevIous four low-order bits of the Accumulator are copied Into the four
high-order bits of the specified memory location.
Suppose the Accumulator contains 7F16' HL register pair contains 400016, and memory location 400016 contains 1216. After execution of the instruction

RRD
the Accumulator will contain 7216 and memory location 400016 will contain F116:
Before
Accumulator
7
\

SE2L
After

Memory

F J
~ rpj}2
" '<, /'
\

High-order blt=:'-S:t :-:

4 ones, set P/O to 1

3-143

Accumulator

Memory

lIDJ

Non-zero result,
set Z to

RST n -

RESTART

S Z ACP/O N

Data
Memory

I I I I:J

B.C
D.E
H.L

SP

ppqq

PC

mmmm

IX

--

mm+1
mm

ppqq-2
ppqq-l
ppqq

{:" ppqq-2 )

.......
~~mmm+y.

IY

1-'

Cooooooooooxxxooo

Program
Memorv

-6

- xxx
11

111

mmmm
mmmm+1
mmmm+2
mmmm+3

lA11 xxx 111

Call the subroutine ongined at the low memory address specified by n.

When the instruction
RST 18H
has executed. the subroutine origlned at memory location 001816 is called. The preVIOUS Program Counter contents are pushed to the top of the stack.
Usually. the RST Instruction is used in conjunction with Interrupt processing. as described in Chapter 12.
If your application does not use all RST instruction codes to service
SUBROUTINE
Interrupts. do not oYerlook the possibility of calling subroutines
CALL USING
uSing RST Instructions. Ongin frequently used subroutines at apRST
propriate RST addresses. and these subroutines can be called with
a single-byte RST instruction instead of a three-byte CALL Instruction.

3-144

SBC A,data-SUBTRACT IMMEDIATE DATA FROM

ACCUMULATOR WITH BORROW

F X

Z AC P/O N [

1X I X 1X 11 'I X I

(.

Data

xx-yy-C ")---

xx

B.C
D.E
H.l

SP
PC

"

mmmm

IX
IV

Memory

DE

I
I

I
R

Program

~mmmm+2

'--

YV

mmmm
mmmm+ 1

mmmm+2
mmmm+3

--

SBC A.

data

DE

-.yy

Subtract the contents of the second object code byte and the Carry status from the Accumulator.
Suppose xx=3A 16 and Carry= 1. After the instruction
SBC A.7CH
has executed, the Accumulator will contain B016
3A
Twos comp of 7C
Twos comp of Carry

1 "" 5" 1
Borrow. set C to 1

0011
1 000
1111
101 1

1010
0 1 00
1111
1 101

fJ
-

Borrow. set AC to 1

1Of 1=0, set Pia to 0

The Carry flag

IS

L Noo-w , '""'t ." Z" 0

Subtract Instruction, set N to 1

set to 1 for a borrow and reset to 0 if there

3-145

IS

no borrow.

SBC A,reg -

SUBTRACT REGISTER WITH BORROW

FROM ACCUMULATOR

.t

Z ACPIQ N C

Ixlxlxlxlqxl
A
B.C
D.E

"
xx

"~c,~I~,
C, D. E, H or L

is yy

H.L

_/'

SP
PC

Data

xx-yy-C

- :i)

~~.J..mmmm+

mmmm

Program

Memorv

IX
IY
I

l00llxxx

mmmm
mmmm+ 1
t---""'1 mmmm + 2
t-----t mmmm + 3

SBC A.

reg

10011

xxx

000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Subtract the contents of the specified register and the Carry status from the Accumulator.
Suppose xx=E316' Register E contains A016. and Carry=l. After the instruction
SBC A.E
has executed. the Accumu lator will contain 4216.
E3
Two's comp of AO
Two's comp of 1

No borrow. set C to 0

1110
0110
1111
0100

0011
0000
1111
0010

Noo -"" re,,'t '" Z ,,0

' - - - - - - N o borrow. set AC to 0

1-+1 =0. set PIO to 0

Subtract Instruction, set N to 1

The Carry flag is set to 1 for a borrow and reset to 0 if there

3-146

IS

no borrow.

SBC A,(HL)SBC A, (IX +disp)

SBC A,(lY+disp)
S

ZACP/ON

SUBTRACT MEMORY AND CARRY FROM

ACCUMULATOR

Data
Memorv

F IXIXIXIXlllX

.:;" .-=C

xx

.1
xx-yy-C) . -

B.C
D.E
H.l

SP
PC
IX
IY
I
R

PP

yy

PP;

qq

mmmm

.-.---::I~mmm + 1

Program
Memorv

~
9E

The illustration shows execution of SBC A.(HL):

mmm
mmm m+ 1
mmmm+2
mmmm+3

SBC A,(HL)
~

9E

Subtract the contents of memory location (specified by the contents of the HL register
paid and the Carry from the Accumulator.
Suppose Carry=O. ppqq=400016' xx=3A16, and memory location 400016 contains
7C 16. After execution of the instruction
SBC A,(HL)
the Accumulator will contain BE16.
3A
Two's comp of 7C
Two's comp of Carry

1 sets S to 1
Borrow, set C to 1

0011
1 000

1010
0100

101 1

1 1 10

f.J

Noo-""

""It '"

Z to

'-------Borrow, set AC to 1

OO=O, set PIO to 0

Subtract instruction, set N to 1

The Carry flag is set to 1 for a borrow and reset to 0 if there IS no borrow.
SBC A,{IX+disp)

"-v-"

-,-

DD 9E

Subtract the contents of memory location (specified by the sum of the contents of the
IX register and the displacement value d) and the Carry from the Accumulator.
SBC A, (IY+disp)

~-,-

FD 9E

This instruction IS identical to the SBC A,{IX+disp) instruction, except that It uses the IY
register Instead of the IX register.

3-147

SBC Hl,rp -

SUBTRACT REGISTER PAIR WITH CARRY

FROM HAND l
SZACPIONC

Data

: I[~[IIx[I1~xI!ExIII1I!ZX}~--------II"'II>1E?
Be. DE. HL or SP
contains ~yyy

B.C

D.E
H.L
SP

I
I

xx

xx

mmmm

PC

-- -

IV
I
R

Program

:1. mmmm + 2

IX

Memorv
ED

mmmm
mmmm + 1
mmmm+2
....---11 mmmm + 3

01xx0010

lX
01

xx 0010

00
01
10
11

for
for
for
for

rp is register pair BC
rp IS register pair DE
rp IS register pair HL
rp IS Stack POinter

Subtract the contents of the designated register pair and the Carry status from the HL
register pair.
Suppose HL contains F4A216. BC contains A03416, and Carry=O. After the instruction
SBC HL,BC
has executed, the HL register pair will contain 546E16:
Two's camp of F4A2
Two's camp of A034
Two's camp of Carry

o sets S to 0
No borrow, set C to 0

1111
0101

0100 1010 0010

1111 1100 1100

0101

0100 0110 1110

tJ

LNon-zero result, set Z to 0

No borrow,

1....

1 1=0, set P/O to 0

The Carry flag

IS

Subtract instruction, set N to 1

set to 1 for a borrow and reset to 0 if there

3-148

IS

no borrow.

SCF -

F'

SET CARRY FLAG

C~

S ZACP/ON

I I I I I

Data

B.C
D.E
H.L

SP
PC

mmmm

:::::;;: '"I mmmm+ 1

""""'-

IX
IY
I
R

Program
emorY

~
37

SCF

mmmm
mmmm+l
mmmm+2
mmmm+3

37
When the SCF instruction IS executed. the Carry status IS set to 1 regardless of ItS preVIOUS value. No other statuses or register contents are affected.

3-149

SET b,reg S

Z AC

SET INDICATED REGISTER BIT

PIO N C

Data

I I I I

S,C
D,E
H,L

VVVV VVVY

SP

PC

".......

I'
mmmm

IX
IY

~:I~mmm+v

Program
Memory

CS
11bbbxxx

mmmm
mmmm + 1
mmmm+2

1----1 mmmm + 3

11\
CB 11 bbb

xxx
-.-

Bit bbb
o 000
1 001
2 010
3 011
4 100
5 101
6 110
7 111

xxx
000
001
010
011
100
101
111

-,-

Register
B

C
D

E
H
L

SET indicated bit within specified register, After the instruction

SET 2.L
has executed. bit 2

In

Register L will be set. (Bit 0

3-150

IS

the least significant bit.)

SET b,(HL) -SET BIT b OF INDICATED MEMORY POSITION

SET b,(IX+disp)
SET b,(IY+disp)
S

Z AC Pia N C

I I I I I I I
yvyy yyyy

P;

B.C
D.E
H,L

qq'

PP

--

SP

PC

mmmm

:>

~~mmm+2

IX
IY
I

Program

Memory

CB
l1bbbllO

mm mm
mm mm+l

I-----f mmmm + 2
I--_ _-fmmmm + 3
The illustration shows execution of SET b,(HU. Bit 0 is the least significant bit

lr~

CB 11 bbb 110
Bit Set

o
1
2
3
4
5
6
7

bbb
000
001
010
011
100
101
110
111

Set Indicated bit within memory location Indicated by HL

Suppose HL contains 400016. After the instruction
SET 5.(HL)
has executed. bit 5 In memory position 400016 will be 1.

DD CB

d 11

bbb 110

bbb is the same as In SET b,(HL)

Set Indicated bit within memory location Indicated by the sum of Index Register IX and
displacement

3-151

Suppose Index Register IX contains 400016. After execution of

SET 6.(lX+5H)
bit 6

In

memory location 400516 will be 1.

FD CB d 11

bbb

110

bbb is the same as

In

SET b. (HU

This instruction IS identical to SET b.(lX+displ. except that It uses the IY register instead
of the IX register.

SLA reg S

SHIFT CONTENTS OF REGISTER LEFT ARITHMETIC

ZAcP/ON

9-

Data

FIX'X'O'X'O' '.
A

/'

",

D.E
H.L
SP
PC
IX
IV

mmmm

")

1-&:""::1 mmmm + 2

......

Program
Memory

CB

001oo1

mmmm
mmmm + 1

1-----1
1-_ _--1

mmmm + 2
mmmm + 3

The illustration shows execution of SLA C.

SLA reg

r~
CB 00100 xxx

-.000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Shift contents of specified register left one bit. resetting the least significant bit to O.
Suppose Register B contains 1F16. and Carry= 1. After execution of
SLA B
Register B will contain 3E16 and Carry will be zero.

3-152

Before

After

Register B

Carry

Register B

Carry

[Q]

10001 1 1 111

~Non-zero result. set Z to 0

o sets S to 0
5 ones. set P/O to 0

SLA (HL) SHIFT CONTENTS OF MEMORY LOCATION

SLA (lX+disp) LEFT ARITHMETIC
SLA (lY+disp)

S
F

Z AC P/O N C

Data
Memory

0X1i1XI I 1411I
0

S.C
D.E
H.L
SP

PC

pp qq
pp

qq

~~mmm+v

mmmm

IX
IY
I

I
I

Program
Memory

mmm m
mmm m+l
mmmm+2
1 - - - - 1 mmmm + 3

CB
26

The illustration shows execution of SLA (HU:

SLA (HU
--.,.-'

CB 26
Shift contents of memory location (specified by the contents of the HL register pairlleft
one bit. resetting the least significant bit to O.
Suppose the HL register pair contains 450016. memory location 450016 contains
8416. and Carry=O. After execution of
SLA (HU
memory location 450016 will contain 0816. and Carry will be 1.
Before

After

Memory

Carry

1100001001

ill

Memory

Carry

ill

o sets S to 0
lone. set P/O to 0

LNon-zero result. set Z to 0

3-153

n
n

SLA (IX+disp)

Shift contents of memory location (specified by the sum of the contents of the IX
register and the displacement value d) left one bit arithmetically. resetting least significant bit to O.
SLA OY+disp)

This instruction
register.

SRA reg S

IS

identical to SLA Ox+displ. but uses the IY register instead of the IX

ARITHMETIC SHifT RIGHT CONTENTS Of

REGISTER

Z AC P/O N C

Data

IXIXIOIXI.OI

i41

S.C

D.E

'-

H.L

SP
PC
IX
IY
I
R

.",.

mmmm

.....

:::;;;::"""'" mmmm+2

~
Program
Memory

CS
00101111

mmmm
mmmm + 1

1I-----lI mmmm + 2
I - - - - i mmmm + 3
The illustration shows execution of SRA A:

--t\

CB 00101 xxx

-.000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Shift specified register right one bit. Most significant bit is unchanged.
Suppose Register H contains 5916. and Carry=O. After the instruction
SRA H
has executed. Register H will contain 2C16 and Carry will be 1.

3-154

Before

After

Register H
10 1 0 1 1 00 1

III

@]

o sets S to 0
3 ones. set P/0 to 0
SRA (HL)SRA (lX+disp)
SRA (lY+disp)

Z ACP!ON

Register H

l.Non-zero result. set Z to 0

ARITHMETIC SHIFT RIGHT CONTENTS OF

MEMORY POSITION

Data
Memory

IX\X\O\X\O\lI
A
B.C
D.E

L
_/ - V

H.L

SP
PC
IX
IY
I

-."'":1 mmmm + 4

mmmm
ppqq

I
I

+d
ppqt

Program
Memory
DO
CB
d
2E

~pqq+d~

mmm m
mmm m+l
mmm m+2
mmm m+3
mmm m+4

The illustration shows execution of SRA (IX+disp):

SRA (IX+disp)
'-v-" -.-

J;:hr
Shift contents of memory location (specified by the sum of the contents of Register IX
and the displacement value d) right. Most significant bit is unchanged.
Suppose Register IX contains 340016. memory location 34AA16 contains 2716. and
Carry=1. After execution of
SRA (IX+OAAH)
memory location 34AA16 will contain 1316. and Carry will be 1.
Before

After

Memory

Carry

1001001111

ill

Memory

Carry

OJ

LNon-zero result. set Z to 0

sets S to 0
3 ones. set P10 to 0

3-155

SRA (IY+disp)

This instruction
register.

IS

J:h

identical to SRA (IX+disp). but uses the IY register Instead of the IX

SRA (HL)
'-v-'

CB 2E
Shift contents of memory location (specified by the contents of the HL register pair)
fight one bit. Most significant bit IS unchanged.

SRL reg -

SHIFT CONTENTS OF REGISTER RIGHT

LOGICAL

r
s

Z AC P/O N

'-

Data

FIOIXIOIXIOI
A
S.C
~.~

H.L
SP
PC
IX
IY
I

...r

-. -:::l mmmm + 2

mmmm

Program

Memorv
mmmm
mmmm + 1
mmmm+2
1----1 mmmm + 3

CS

00111011

The illustration shows execution of SRL E:

l\

CB 00111 xxx

000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Shift contents of specified register fight one bit. Most significant bit
Suppose Register D contains 1F16. and Carrv=O. After execution of

SRL D
Register D will contain OF16. and Carry will be 1.

3-156

IS

reset to O.

Before

After

Register D

Carry

Register D

Carry

1000111111

@]

10000 11111

IJJ

L...Non-zero result, set Z to 0

4 ones. set P/0 to 1

SRL (HL)SRL (IX+disp)

SRL (lY+disp)

SHIFT CONTENTS OF MEMORY LOCATION

RIGHT LOGICAL

Z AC Pia N C

: m:EJ 0 I x ro:~~
B,C
D,E
H,L

sp

PC

pp

qq

l..OiL.~mmm+V

mmmm

IX
IY
I

Program
Memory

mmm m
mmmm+l
I -_ _-immmm + 2
I -_ _-immmm + 3
CB

3E

The illustration shows execution of SRL (HL):

SRL (HL)
~

CB 3E
Shift contents of memory location (specified by the contents of the HL register paid
right one bit Most significant bit is reset to 0,
Suppose the HL register pair contains 200016. memory location 200016 containS 8F16.
and Carry=O, After execution of
SRL (HL)
memory location 200016 will contain 4716. and Carry will be 1.
Before

After

Memory

Carry

Memory

Carry

1100011111

[Q]

1010001111

OJ

L.

4 ones. set P/0 to 1

Non-zero result, set Z to 0

SRL (IX+disp)

Shift contents of memory location (specified by the sum of the contents of the IX
register and the displacement value d) right one bit Most Significant bit IS reset to 0,

3-157

SRL (IY+disp)

J:h,
This instruction is identical to SRL (IX+disp), but uses the IY register Instead of the IX
register.

SUB data-SUBTRACT IMMEDIATE FROM ACCUMULATOR

S

ZACP/ON

[x Lx JxIX 11 I xI

B.C

Data

xx

~~:x:yyr

D.E

H.L

SP

~~mmm+)

mmmm

PC
IX
IY

I
R

Program
Memory

yy

D6
I-

mmmm
mmmm+l
mmmm+2

mmmm+3

SUB

data

D6

VV

Subtract the contents of the second oblect code byte from the Accumulator.
Suppose xx=3A16 .After the instruction
SUB 7CH
has executed. the Accumulator will contain BE16.
3A
001 1
Two's comp of 7C = 1 000
1 01 1
1 sets S to 1
Borrow, set C to 1

o OdO,

1010
01 00
1 1 10

U
t L

NOO' " ' ' co,"". '" Z to 0

Borrow. set AC to 1

set P10 to 0

Subtract Instruction, set N to 1

Notice that the resulting carry is complemented.

3-158

SUB reg -

SUBTRACT REGISTER FROM ACCUMULATOR

S Z ACP/0N C

~
A

xx

B.C
D.E

r.

-,

.......

}-coo"IA'

c,

D. E. H or L IS YY

H.L

SP
PC
IX
IY

Data

xx-yy

mrnmm

::::;;0:--::1 mmmm + 1

......

~
Program
Memory

mmmm
_ _-Immmm + 1
~_ _-Immmm + 2
~_ _-Immmm + 3
10010xxx

SUB

reg

'-".-'

10010

XXX

000
001
010
011
100
101
111

for
for
for
for
for
for
for

reg=B
reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Subtract the contents of the specified register from the Accumulator.

Suppose xx=E3 and Register H contains A016. After execution of
SUB H
the Accumulator will contain 4316.
E3
1 1 10
Two's comp of AO = 01 1 0
0100

o sets S to 0
No borrow, set C to 0

001 1
0000
0011

LNo"""m re'Olt '" Z to

'-------No borrow. set AC toO

Subtract Instruction, set N to 1

1 1=0, set P/O to 0

Notice that the resulting carry is complemented.

3-159

SUB (HL) - SUBTRACT MEMORY FROM ACCUMULATOR

SUB (IX+disp)
SUB (IY+disp)
S Z AC Pia N C

'xlXIXlxlllxl

- ).-

"

xx-yy

'-.--:1

xx

Data
Memorv
yy

ppqq +d

B.C
D,E
H.L
SP

PC
IX

--

,.... ~mmm+v

mmmm
ppqq

IV

I
R

~pqq+d~

Program
Memory

DD
96
d

mmm m
mmm m+ 1

mmmm+2
mmm m+3

The illustration shows execution of SUB (lX+d):

SUB (lX+disp)

~-.-

DD 96

Subtract contents of memory location (specified by the sum of the contents of the IX
register and the displacement value d) from the Accumulator.
Suppose ppqq=400016. xx=FF16. and memory location 40FF16 contains 5016. After
execution of
SUB (lX+OFFH)
the Accumulator will contain AF16
FF

Two's comp of 50 =

1 sets S to 1

1 111
10 11
1 01 0

1111
0000
1111

fJ

LNoo."ro re,"1t .., Z to a

.....- - - - - No borrow. set AC to 0

No borrow. set C to 0

Subtract instruction, set N to 1

1 "'1-1 =0. set PIO to 0

Notice that the resulting carry

IS

complemented,
SUB (lY+disp)
---...-..
-.FD 96

This Instruction IS identical to SUB (IX+displ. except that It uses the IY register Instead
of the IX register.
SUB {HU
---...-..
96
Subtract contents of memory location (specified by the contents of the HL register pair)
from the Accumulator,

3-160

XOR data S

EXCLUSIVE-OR IMMEDIATE WITH ACCUMULATOR

Z ACP/O N C

Data

F~
A

;,,;,,&

XX

B.C
D.E

~~mmm+~

H.L

SP
PC
IX

~x-;;'yyr

mmmrn

~
Program
Memorv

IY
I

EE

'-

XOR

data

EE

Vy

Vy

mmmm
mmmm+ 1
mmmm+2
mmmm+3

Exclusive-OR the contents of the second oblect code byte with the Accumulator.
Suppose xx=3A16. After the instruction
XOR 7CH
has executed. the Accumulator will contain 4616.
3A
001 1
7C = 0 1 1 1
0100

o sets S to

1010
1 1 00
0110

0..J

LNon-zero result. set Z to 0

LThree 1 bits. set PIO to 0

The Exclusive-OR instruction

IS

used to test for changes

3-161

In

bit status.

XOR reg S

EXCLUSIVE-OR REGISTER WITH ACCUMULATOR

Z ACP/O N C

f,

IXIXI1 iXIOIOI
A

B,C
D,E
H,L

-.

::1....... xx "f yy)

~Coo.~ol~'
D, E, H or L
C,

IS yy

SP

PC

/"

...--,. mmmm + 1

mmmm

IX
IY

I
R

Data

~
Program

Memory

I
I

10101xxx

mmmm
mmmm+l

I I - - - - I mmmm + 2
II-_ _-Immmm + 3

XOR

reg

10101

xxx

000 for reg=B

001
010
011
100
101
111

for
for
for
for
for
for

reg=C
reg=D
reg=E
reg=H
reg=L
reg=A

Exclusive-OR the contents of the specified register with the Accumulator,

Suppose xx=E316 and Register E contains A016, After the instruction
XOR E
has executed, the Accumulator will contain 4316,
E3
AO =

1 1 10
10 10
0100

001 1
0000
0011

o sets S to 0....J

Non-zero resu It. set Z to 0

LThree 1 bitS, set PIO to 0

The Exclusive-OR instruction IS used to test for changes

3-162

In

bit status.

XOR (HL) - EXCLUSIVE-OR MEMORY WITH ACCUMULATOR

XOR (IX +disp)
XOR (lY +disp)
S

Z AC PIO N C

Ixlxl 1 1XIOIOI

-..---:t'" xx yy ~

xx

S.C

Data
Memory

yy

ppqq +d

o.E

......-:.J., mmmm+ 3

H.L

SP

PC

mmmm

IX

ppqq

IY
I

"

10-

le..(ppqq+~

Program
Memory

00

AE
d

mmm m
mmm m+l
mm mm+2
mm mm+3

The illustration shows execution of XOR (IX+disp):

XOR (IX+disp)
' - v - -.-

DD AE

Exclusive-OR contents of memory location (specified by the sum of the contents of the
IX register and the displacement value d) with the Accumulator.
Suppose xx=E316. ppqq=450016. and memory location 45FF16 contains A016. After
the Instruction
XOR (IX+OFFH)
has executed. the Accumulator will contain 4316.
E3
AO =

1 1 10

001 1

10 10

0000

0100

0011

o..~ "o..J

LNOO-"" co,,1L .., Z ,,0

LThree 1 bits. set P/O to 0

XOR (IY+disp)
~-.-

FD AE

This instruction is identical to XOR (IX+displ. except that It uses the IY register Instead
of the IX register.
XOR (HU
--..-'

AE
Exclusive-OR contents of memory location (specified by the contents of the HL register
paid with the Accumulator.

3-163

8080A/Z80 COMPATIBILITY
Although the zao microprocessor can certainly be used on
its own merits. one of its important characteristics is its
compatibility with the aOaOA microprocessor. This compatibility has the following features:
1)
2)
3)
4)

8080A/Z80

COMPATIBILITY
FEATURES

All 8080A machine language Instructions are also Z80 machine language instructions.
All 8080A registers are also Z80 registers (see Table 3-6).
Almost all 8080A programs will run on a Z80. with some minor differences to be
noted later.
The Z80 has instructions, registers, and other features not present on the 8080A.
so Z80 programs will not generally run on 8080A processors.

Note that this compatibility does not extend to assembly

language source statements since Z80 assemblers and 8080A
assemblers use different operation code mnemonics. Table 3-7
contains a list of the aOaOA mnemonic codes and the corresponding Z80 codes. while Table 3-8 is the same list organized
by Z80 codes.

8080AlZ80
ASSEMBLY
LEVEL
CONVERSION

Readers should note the binary coding limitations that thiS compatibility places on the extra featu res of the Z80 microprocessor.
The 8080A has some unused operation codes (see Table 3-9) that
are used for some of the Z80's extra Instructions. But there are
simply not enough such codes to cover the large number of
features in a simple form.

8080A
UNUSED
OPERATION
CODES

Thus, many of the added Z80 instructions require a 2-byte operation code. The first byte IS CB, DO, ED. or FD. Note the follOWing
meanings of these codes from Table 3-9:

2-BYTE
OPERATION
CODES

CB DO ED FD -

a register or bit operation

an operation involVing register IX
a miscellaneous non-8080A instruction not covered elsewhere
an operation involVing register IY

The second byte of the operation code describes the actual operation to be performed.
FASTER AND
The end result is that these multi-byte instructions execute rather
slowly (and use more memory) because an additional memory
SLOWER
access is required. The reader should be aware of this vanatlon In
EXECUTING
execution times and try to use faster executing instructions when
INSTRUCTIONS
possible. This warning particularly applies to the extra shift
instructions (RLC, RRC, RL. RR. SRA, SRL) and to instructions Involving the Index
registers IX and IY.
There are a few minor incompatibilities between the
8080A and the Z80. These are:
1)

2)

3)

8080A/Z80
INCOMPATIBILITIES

The Z80 uses the P (or P/O) flag to Indicate twos complement overflow after anthmetlc operations. The 8080A always uses thiS flag for
panty.
The Z80 and 8080A execute the DAA Instruction differently. On the Z80, thiS instruction will correct decimal subtraction as well as decimal addition. On the
8080A. It will correct only decimal addition.
The Z80 rotate instructions clear the AC flag. The 8080A rotate Instructions do
not affect the AC flag.

3-164

Table 3-6. Register and Flag Correspondence between

Z80 and 8080A
zao Register

SOaOA Register

A
A'

None

B
None

B'
C
C'

C
None

0
0'

None

E
E'

None

Least Significant Half of PSW

None

H'

None

None

IX

None

IY

None

L'

None

None

PC

PC

SP

SP

zao Register Pelrs

BOaOA Register Pairs

BC

DE

HL

AF

PSW

zao Flags

aOaOA Flags

C (Carry)

C (Carryi

H (Half-Carry)

AC (Auxiliary Carryi

N (Subtract)

None

Pia (Paritv/Overflowl
S(Sign)

P (Paritvi
S(Sign)

Z (Zero)

Z (Zeroi

The Z80 is not compatible with the extra features of

the 8086 microprocessor. The codes used for RIM and
SIM on the 8085 are used for relative jumps (NZ and NC) on
the Z80.

8085/Z80
INCOMPATIBILITIES

Instruction timings on the 8080A, 8085, and Z80 all

differ. Programs that depend on precise instruction timings will therefore execute properly only on the processor for which they were written.

TIMING
INCOMPATIBILITIES

The N flag on the Z80 occupies bit 2 of the F register; the corresponding bit in the
Processor Status Word of the 8080A is always a logic '1'.

3-165

Table 3-7 Correspondence between 8080A and Z80 Mnemonics

aoaOA

zao Mnemonic

aoaOA Mnemonic

zao Mnemonic

Mnemonic

HL,(addr)

ADC

A,data

LHLD

addr

LD

ADC

A,rag or (HLI

LXI

rp,datale

LD

rp,datale

reg or M

ADD

A,reg or (HU

MOV

reg,reg or

LD

reg,rag or (HU

ADD

A,data

MOV

reg or M,reg

LD

rag or (HLI,reg

AND

reg or (HU

MVI

reg or M,data

LD

reg or (HU,data

ACI

data

ADC

reg or

ADD
ADI

data

ANA

reg or

NOP

ANI

dete

AND

data

NOP

CALL

addr

CALL

eddr

ORA

reg or

CC

addr

CALL

C.addr

ORI

data

OR

data

CM

addr

CALL

M,addr

OLrr

port

OUT

(portl.A

CMA

CPL

PCHL

CMC

CCF

POP

pr
pr

OR

reg or (HLI

JP

(HU

POP

pr

PUSH

pr

CMP

reg or M

CP

reg or (HU

PUSH

CNC
CNZ

addr
addr

CALL
CALL

NC,addr
NZ,addr

RAL
RAR

CP

addr
addr

CALL

P,addr

RC

RET

CPE

CALL

PE,addr

RET

RET
RLCA

RLA
RRA
C

CPI

data

CP

data

RLC

CPO

addr

CALL

PO,addr

RM

RET

CZ

addr

CALL

Z,addr

RNC

RET

NC

rp

ADD

HL,rp

DAA
DAD

DAA

RNZ

RET

NZ

RP

RET

OCR

reg or M

DEC

reg or (HLI

RPE

RET

PE

DCX

rp

DEC

rp

RPO

RET

PO

01

RRC

RRCA

EI

EI

RST

HLT

HALT

RZ

01

n
M

RST

RET

sac

IN

A,(portl

saa

reg or

INC

rag or (HU

SBI

data

SBC

A.reg or (HLI
A,data

rp

INC

rp

SHLD

addr

LD

(addrl.HL

JP

LD

JP

C.addr
M,addr

SPHL

JM

addr
addr

STA

LD

SP,HL
(addr).A

LD

(BCI or (DEI.A

IN

port

INR

reg or

INX
JC

addr
8 or 0

JMP

addr

JP

addr

STAX

JNC

addr

JP

NC,addr

STC

JP

addr

JP

P,addr

SUB

reg or M

SUB

JNZ

addr

JP

NZ,addr

SUI

data

SUB

data

JPE
JPO

JP
JP

PE,addr
PO.addr

XCHG
XRA

reg or M

EX
XOR

DE,HL
reg or (HLl

JZ

addr
addr
addr

JP

Z,addr

XRI

data

XOR

LOA

addr

LD

A,(addrl

XTHL

data
(SPI.HL

LDAX

a or 0

LD

A,(aC) or (DEI

3-166

SCF

EX

rag or (HU

Table 3-8. Correspondence between l80 and 8080A Mnemonlc~

zao

aOaOA Mnemonic

Mnemonic

Mnemonic

BOaOA

A,data

ACI

data

INC

rp

INX

ADC

A(HLI

ADC

INC

reg

INC

xv
(xv + disp)

Mhemonic
rp

ADC

Areg

ADC

ADC

A.lxv + displ

ADC

HL,rp

ADD

A,data

ADI

data

ADD

A,(HLI

ADD

INIR

ADD

A.reg

ADD

reg

JP

addr

JMP

addr

ADD

JP

C,addr

JC

addr

ADD

A(xv + displ
HL,rp

JP

(HLI

PCHL

ADD

IX,pp

JP

M.addr

JM

addr

ADD

IY.IT

JP

NC.addr

JNC

bAD

IND

INDR

INI

rp

AND

data

ANI

data

JP

NZ,addr

JNZ

addr
addr

AND

(HLI

ANA

JP

P,addr

JP

addr

AND

reg

ANA

reg

JP

PE.addr

addr

AND

(xv + disp)

JPE

JP

PO,addr

JPO

addr

BIT

b,(HLI

JP

Z.addr

JZ

addr

BIT

b,reg

JP

xv

BIT

b,{xy + displ
addr

JR

CALL

C.disp
disp

CALL
CALL

CALL

addr

JR

C.addr

CC

JR

NC,disp

M.addr

CM

addr
addr

JR

NZ.disp

CALL

NC,addr

CNC

addr

JR

Z,disp

CALL

NZ.addr

CNZ

addr

LD

A.laddrl

LDA
LDAX

CALL

P.addr

CP

addr

LD

A.lBC) or (DE)

CALL

PE,addr

CPE

addr

LD

AI

CALL

PO,addr

CPO

addr

LD

Z,addr

CZ

addr

LD

AR
(addr),A

CALL

CCF

CMC

CP

CPI

LD

(addr),BC or DE

data

STA

addr
8 or 0

addr

data
(HL)

LD

(addrl,HL

SHLD

CMP

LD

(addrl,SP

CP

reg

CMP

reg

LD

(addrl,xv

(xv + displ

CP

LD

(BCI or (DE),A

STAX

LD

BC or DE,(addr)

LD

HL,(addr)

LHLD

LD

(HLl,data

MVI

M,dats

LD

(HLI,reg

MOV

M,reg

CPL

CMA

LD

I,A

DAA

DAA

LD

R,A

CP

CPD
CPDR
CPI
CPIR

DEC

(HLI

DEC

reg

DEC
DEC

rp

DEC

{XV + dispj

xv

DI
DJNZ

disp

EI

DCR

LD

reg,data

DCR

reg

LD

reg.lHL)

DCX

rp

LD
LD

reg,reg

reg.(xv + dispi

reg,M
reg,reg

LD

rp,datat6

LXI

SP,(addri

LD

SP,HL

SPHL

EI

LD

SP,xv

EX

AF,AF

LD

xY,data16

DE,HL

XCHG

LD

xy.laddrl

EX

(SP),HL

XTHL

LD

(xv + dispi,dat8

EX

(SP),xv

LD

(xy

EXX

LDD

HALT

HLT

LODR

LOI

IN

A.lpOrtl

IN

IN

reg.lCI

INC

(HLI

INC

reg

rp,datat6

port

LOIR

INR

NEG
Nap

INR

reg

OR

3-167

addr

reg,data

LD

indicates that there IS no corresponding lnstructlon.

8 or 0

MVI

DI

+ disp),reg

addr

MOV
MOV

EX

1M

zao

ADC

Nap

data

ORI

data

Table 3-8. Correspondence between Z80 and 8080A Mnemonics (Continued)

Z80 Mnemonic

8080A Mnemonic

Z80 Mnemonic

8080A Mnemonic

OR

IHLI

ORA

RR

IHLI

OR

reg

ORA

reg

RR

reg

OR
OTOR

Ixv + displ

Ixv + displ

RR
RRA

RRC

IHLI

RRC

reg

RRC

(xv + disp)

OTIR
OUT

ICl,reg

OUT

IportlA

OUT

port

OUTO
OUTI

RRCA

RRC

RRO

RST

RST

S8C

A,data

581

data

S8C

A,IHLI

SB8

SBC

A,reg

588

reg

SBC

A,(xv + displ

SBC

HL,rp

SCF

RET

SET

b,(HLI
b,reg
b,(xv + disp)

POP

pr

POP

POP

xv

PUSH

pr

PUSH

PUSH
RES

xv
b,lHLI

RES

b,reg

RES

b,lxv + displ

RET

RAR

pr
pr

STC

RET

RC

SET

RET

RM

SET

RlIT

NC

RNC

SLA

RET

NZ

RNZ

SLA

IHLI
reg

RET

RP

SLA

(XV

IHLI

+ displ

RET

PE

RPE

SRA

RET

PO

RPO

SRA

reg

RET

RZ

SRA

(xv + displ

SRL

SRL
SRL

IHL)
reg

RETI
RETN
RL

(HLI

RL

reg

RL

(XV

+ displ

RLA
RLC

IHLI

RLC

reg

RLC

(xv + disp)

RLCA

IS

SU8

data

SUI

data

SU8

IHLI

SU8

RAL

SU8

reg

SUB

reg

SUB

Ixv + displ

XOR

data

XRI

data

XOR

XRA

RLC

XOR

IHLI
reg

M
reg

XOR

(xv + displ

RLO
indicates that there

no corresponding instruction

3-168

(xv + dispi

XRA

Table 3-9. Unused 8080A Operation Codes and Their Z80 Meanings
8080A Operation Code
OB

Z80 Use
EX AF.AF'

10

DJN7 disp

18

JR disp

20 IRIM on 8085i

JR NZ.disp

28

JR Z.disp

30 ISIM on BOB5)

JR NC.disp

3B

JR C.disp

CB

BIT. RES. RL. RLC. RR. RRC. SET. SLA. SRA. SRL

OS

EXX

DO

All instructIons involving RegIster IX.

ED

ADC

HL.rp

A.I

LD

A.R

OTDR

CPDR

LD

laddr).rp

OTIR

CPI

LD

I.A

OUT

CPIR

LD

R.A

aUTO

rp,(addr)

OUTI

1M

LD

IN

reg.IC)

LDD

RETI

IND

LDDR

RETN

INDR

LDI

RLD

INI

LDIR

All InstructIons Involvmg RegIster IY.

3-169

ICl.res

RRD
SBC

INIR
FD

NEG

LD

CPO

HL.rp

ZILOG

zao ASSEMBLER CONVENTIONS

The standard zao assembler is available from zao manufacturers and on the major
time-sharing networks; it is also part of most development systems. Cross assembler versions are available for most large computers and many minicomputers.

ASSEMBLER FIELD STRUCTURE

The assembly language instructions have the standard field structure (see Table
2-1). The required delimiters are:
1)
2)
3)
4)
5)

A colon after a label, except for the pseudo-operations EQU, DEFL, and
MACRO, which require a space.
A space after the operation code.
A comma between operands in the operand field. (Remember this oneil
A semicolon before a comment.
Parentheses around memory references.

TYPical
START

zao assembly language Instructions are:

LD
ADD
HALT

A.(1000l
HL.DE

:GET LENGTH

LABELS
The assembler allows six characters in labels; the first character must be a letter,
while subsequent characters must be letters, numbers, ?, or the underbar
character (_I. We will use only capital letters or numbers, although some versions
of the assembler allow lower-case letters and other symbols.

RESERVED NAMES
Some names are reserved as keywords and should not be used by the programmer. These are the register names (A, B, C. D. E, H, L, I, Rl. the double register
names (IX, IV, SPI. the register names (AF, BC. DE. HL, AF'. BC', DE', HL'I. and
the states of the four testable flags (C, NC, Z, NZ, M, P, PE. PO).

PSEUDO-OPERATIONS
The assembler has the following basic pseudo-operations:

DEFB

DEFINE BYTE

DEFL
DEFM
DEFS
DEFW
END
EQU
ORG

DEFINE LABEL
DEFINE STRING
DEFINE STORAGE
DEFINE WORD
END
EQUATE
ORIGIN

DEFB. DEFM. and DEFW are the Data pseudo-operations used to

DEFB,DEFM,
place data in ROM. DEFB IS used for 8-blt data. DEFW for 16-blt
DEFW
data. and DEFM for ASCII strings (63 or less characters long). The
PSEUDOonly unusual feature to remember IS that DEFW stores the eight
OPERATIONS
least significant bits of data In the first word and the eight most
significant bits In the second word. This IS the standard 8080A/8085/Z80 procedure for
stonng addresses In memory. but IS contrary to normal practice. You must be aware of
the order when stonng 16-blt data.

3-170

Note that DEFB and DEFW define the value of only a single byte or single word. respectively. Establishing a table of values requires a series of DEFB or DEFW pseudo-operations. one for each byte or word of data.

Examples:
ADDR:
DEFW
3165H
results in (ADDR) = 65. and (ADDR+1) = 31 (hexadecimal).
TCONV' DEFB
32
ThiS pseudo-operation places the number 32 In the next byte of ROM and assigns the
name TCONV to the address of that byte.
ERROR: DEFM
'ERROR'
ThiS pseudo-operation places the 7-bit ASCII characters E. R. R. O. and R In the next five
bytes of ROM and assigns the name ERROR to the address of the first byte.
OPERS:

DEFW
DEFW
DEFW
DEFW

FADD
FSUB
FMUL
FDIV

ThiS series of pseudo-operations places the addresses FADD. FSUB. FMUL. and FDIV In
the next eight bytes of memory and assigns the name OPERS to the address of the first
byte. Note that the first byte contains the least significant bits of address FADD.
DEFS IS the Reserve pseudo-operation used to assign locations
RAM: It allocates a specified number of bytes.

In

EQU IS the Equate or Define pseudo-operation used to assign

values to names.
DEFL IS similar to EQU. except that DEFL allows the name to be
redefined later. DEFL IS much like the SET directive In other assemblers. It should only be used to define assembly time variables
(i.e.. those variables used In conditional assembly or conditional
macro expansion statements).
ORG IS the standard Origin pseudo-operation.

zao programs usually have several origins: the origins are used as
follows:

DEFS
PSEUDOOPERATION
EQU
PSEUDOOPERATION
DEFL
PSEUDOOPERATION

ORG
PSEUDOOPERATION

1)
2)

To specify the RESET address (usually zero).

To specify interrupt entry POints (usually 0 to 6616 but may be anywhere in memory.

3)
4)
5)
6)
7)

To
To
To
To
To

specify the starting address of the main program.

specify the starting addresses of subroutines.
define areas for RAM storage.
define an area for the RAM Stack.
specify addresses used for I/O ports and special functions.

3-171

Examples:
RESET

EOU
0
ORG
RESET
This sequence places the RESET instruction sequence in memory beginning at address

O.
INT1

EOU
ORG

38H
INT1

The instruction sequence that follows is stored

In

memory beginning at location 3816'

END simply marks the end of the assembly language program.

The special purpose pseudo-operations CONDo MACRO, ENDC.
and ENDM are described later in this chapter.

END

~~i~~~ioN

LABELS WITH PSEUDO-OPERATIONS

The rules and recommendations for labels with
follows:
1)
2)
3)

zao

pseudo-operations are as

EOU, DEFL, and MACRO require labels, since the function of these pseudo-operations IS to define the meaning of that label.
DEF8, DEFM. DEFW, and DEFS usually have labels.
ORG, CONDo EN DC. ENDM, and END should not have labels, since the meaning of
such labels IS unclear.

ADDRESSES
The Zilog Z80 assembler allows entries
of the following forms:

any

NUMBERS AND
CHARACTERS
IN ADDRESS
FIELD

All anthmetic and logic operations within an address field assume

all arguments are 16-bit data: they produce 16-bit results, These
operations are allowed as part of expressions In the address field.

ASSEMBLER
ARITHMETIC
AND LOGICAL
OPERATIONS

In

the address field

In

1)

Decimal (the default case)

Example: 1247

2)

Hexadecimal (must start with a digit and end with an H)

Examples: 142CH. OE7H
Octal (must end with a or 0, but 0 IS far less confusing)
Example: 12470 or 12470
Binary (must end with B)
Example: 1001001000111B

3)
4)
5)

ASCII (enclosed In single quotation marks)

Example: 'HERE'

6)

As an offset from the Program Counter ($)

Example: $+237H

When defining address constants, hexadecimal notation should

be used. Binary constants of 16 bits are unwieldy and hence errorprone. Octal constants are inconvenient due to the fact that addresses are stored in
low-order byte high-order byte format. ThiS diVISion occurs In the middle of an octal
digit. which causes you to have to split a digit. For example. to express the address
9D7FH or 1165770 In low-high format you get 7F9DH or 772360. As you can see. In
hexadecimal notation the digits are simply transposed. while no such simple relationship eXists for octal notation.

3-172

OPERATOR

.NOT. or \
.RES.

...
/

.MOD.
.SHR.
.SHL.

+
-

.AND. or &
.OR. or I
.XOR.
.EO. or =
.GT. or >
.LT. or <
.UGT.
.ULT.

FUNCTION

PRIORITY

UNARY PLUS
UNARY MINUS
LOGICAL NOT
RESULT
EXPONENTIATION
MULTIPLICATION
DIVISION
MODULO
LOGICAL SHIFT RIGHT
LOGICAL SHIFT LEFT
ADDITION
SUBTRACTION
LOGICAL AND
LOGICAL OR
LOGICAL XOR
EQUALS
GREATER THAN
LESS THAN
UNSIGNED GREATER THAN
UNSIGNED LESS THAN

1
1
1
1
2

3
3
3
3
3
4
4
5
6
6
7
7
7
7
7

In address expressions with more than one operator. the order of evaluation IS defined
by the priorities given in the list above. Operators having the same priorities are evaluated from left to right. Expressions In parentheses are evaluated first. Remember that
enclosing an expression entirely In parentheses Indicates a memory address.

Note the following:

1)
2)

The Result operator (.RES.) causes overflow to be suppressed; I.e.. a change in sign
caused by overflow into the sign bit does not result In an assembler error.
The shifts have the form:
.SHR.
.SHL.

op 1.0p2
op 1. op2

where op1 IS the number to be shifted and op2 IS the number of shifts. The shifts
are logical. I.e.. zeros are shifted Into the high-order or low-order bits. respectively.
3)

The comparison operators produce a result of either logical True (all ones) or logical
False (zero).

4)

The operators .GT. and .LT. assume signed twos complement numbers. whereas
.UGT. and .ULT. assume unsigned operands. This means that. for .GT. and .L T..
positive twos complement numbers are larger than negative twos complement
numbers. while the opposite IS the case for .UGT. and .ULT.

3-173

CONDITIONAL ASSEMBLY
The ZSO assembler has a simple conditional assembly
capability based on the pseudo-operations COND and ENDC.
COND IS followed by an expression. for example:
COND
COND

BASE - 1000H
or
BASE - OPER1

COND AND
ENDC
PSEUDOOPERATIONS

If the expression is not zero. the assembler Includes all of the instructions up to the
ENDC pseudo-operation In the program: if the expression IS zero. the assembler Ignores
all instructions between COND and ENDC.
We will not use conditional assemblies or refer to this capability again: It IS sometimes
handy for adding or eliminating debugging instructions. or configuflng unique versions
of a common program.

MACROS
The standard ZSO assembler has a macro capability that
assigns names to instruction sequences. Use the pseudo-operation MACRO to begin the definition and ENDM to end It. The
macro may have parameters and may Include any assembly
language Instructl0!1s except the definitions of other macros.

MACRO AND
ENDM
PSEUDO
OPERATIONS

The macro capability IS often a convenient programming shorthand. but we will not use
It.
Note that instruction sequences defined by macros are generally qUite short they
should not exceed ten or fifteen instructions. longer sequences Should be made Into
subroutines to conserve memory space.
Every MACRO pseudo-operation must have a label: the label is the name with which
you Identify the macro. For a diSCUSSion of thiS sublect. see Chapter 2.

3-174

Chapter 4
SIMPLE PROGRAMS
The only way to learn assembly language programming is to write assembly
language programs. That is whilt we will do for the next six chapters. which contain examples of typical microprocessor tasks. Problems at the end of each
chapter contain variations on the examples given in the text of the chapter. You
should try to run the examples on a ZSO-based microcomputer system to ensure
that you understand the material covered in the chapter.
In thiS chapter we begin with some very simple programs.

GENERAL FORMAT OF EXAMPLES

Each program example contains the following parts:
1)

A title that describes the general problem.

2)

A statement of purpose which describes the specific task that

the program performs. plus the memory locations that it uses.

3)

A sample problem showing input data and results.

4)

A flowchart if the program logic IS complex.

5)

The source program or assembly language listing of the program.

6)

The object program or hexadecimal machine lang.uage listing of the program.

7)

Explanatory notes that discuss the instructions and methods used in the program.

The problems at the end of the chapter are similar to the examples; problems
should be programmed on a ZSO-based microcomputer system using the examples
as guidelines.
The source programs in the examples have been constructed as follows:

...----....,
GUIDELINES
FOR
EXAMPLES

1)

Standard Zilog Z80 assembler notation IS used. as summarized In Chapter 3.

2)

The forms In which data and addresses appear are selected for
clarity rather than for consistency. We use hexadecimal numbers for memory addresses. instruction codes. and 8CD data: decimal for numeric
constants: binary for logical masks: and ASCII for characters.

3)

Frequently used instructions and programming techniques are emphasized.

4)

Examples illustrate tasks that microprocessors perform In communications. instrumentation. computer. business equipment. industrial. and military applications.

5)

Detailed comments are included.

6)

Simple and clear structures are emphaSized. but programs are as effiCient as possible within thiS gUideline. The notes often describe more efficient procedures.

7)

Programs use consistent memory allocations. Each program starts in memory location 0000 (the RESET location) and ends With the HALT instruction. If your

4-1

microcomputer has no monitor and no Interrupts. you may prefer to end programs
with an endless loop instruction. e.g.:
HERE:

JR

HERE

The hexadecimal version IS 18 followed by FE. You may replace the HALT or JR
HERE instruction with a RESTART or JP instruction that transfers control back to
the monitor in some Z80-based microcomputers.
Consult the user's manual for your microcomputer to determine the required memory
allocations and terminating instruction for your particular system.

GUIDELINES FOR PROBLEMS

When tackling the problems at the end of each chapter. try
to work within the following guidelines:

PROGRAMMING
GUIDELINES

1)

Comment each program so that others can understand it.

The comments can be bnef and ungrammatical: thev
should explain the purpose of a section or instruction in the program. Comments
should not describe the operation of Instructions: that description is available In
manuals. You do not have to comment each statement or explain the obvious. You
may follow the format of the examples but provide less detail.

2)

Emphasize clarity. simplicity. and good structure in programs. While programs

should be reasonably efficient do not worry about saving a single byte of program
memory or a few microseconds.

3)

Make programs reasonably general. Do not confuse parameters (such as the number of elements in an array) with fixed constants (such as 7T or ASCII C).

4)

Never assume fixed initial values for parameters. i.e.. use an instruction to load an
Initial value into a parameter.

5)

Use assembler notation as shown In the examples and defined in Chapter 3.

6)

Use hexadecimal notation for addresses. Use the clearest possible form for data.

7)

If your microcomputer allows it start all programs In memory location 0000 and
use memory locations starting with 0040 for data and temporary storage. Other16
wise. establish equivalent addresses for your microcomputer and use them consistently. Again. consult the user's manual.

8)

Use meaningful names for labels and vanables. e.g.. SUM or CHECK rather than X.

Y. or Z.
9)

Execute each program on your microcomputer. There is no other way of ensuring

that your program is correct. We have provided sample data with each problem. Be
sure that the program works for special cases.

We now summarize some useful information that you should keep in mind when
writing programs.
USING THE
Almost all processing instructions (e.g.. ADD. SUBTRACT.
ACCUMULATOR
AND, OR) use the Accumulator. In most cases you will load
data Into the Accumulator With LD. using either LD A, (addrl to
load data from any memory location or using LD A.(HU to load
data from the address specified in Registers Hand L. Remember that the parentheses
indicate a memory address rather than data.

The preferred method of accessing memory is using implied addreSSing via Registers Hand L. that IS. uSing (HL). This code causes
the Z80 to perform a memory access uSing the address stored in
Registers Hand L. You can use LD HL. data 16 to load a fixed num-

4-2

USING
REGISTER
PAIR HL

ber Into Registers Hand L or LD HL,(addrl to load the contents of two successive memory locations Into Hand L. You can use INC HL or DEC HL to increment or decrement (by
1) the address in Registers Hand L.
The 8-bit anthmetic and logical operations all use the data In the Accumulator as one of
their operands and place their result into the Accumulator.
Some of the 8-bit anthmetlc and logical operations have special
uses, for example:

SPECIAL
INSTRUCTIONS

SUB A (or XOR A) clears the Accumulator.

ADD A,A shifts the Accumulator left one bit logically. This instruction also multiplies
the contents of the Accumulator by 2. AND A (or OR A) clears the Carry flag while
preserving the contents of the Accumulator.
A logical AND can mask off parts of a word. The required mask has T bits In the pOSItions that you want to reserve and '0' bits In the positions that you want to clear.

PROGRAM EXAMPLES
Ones Complement
Purpose: Logically complement the contents of memory location 0040 and place thA
result into memory location 0041.
Sample Problem:

Result:

(0040)

6A

(0041)

95

Source Program:

LD
CPL
LD
HALT

A,(40H)
(41Hl.A

:GET DATA
;COMPLEMENT
:STORE RESULT

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007

3A
40
00
2F
32
41
00
76

Instruction
(Mnemonic!
LD

A,(40H)

CPL
LD

(41Hl.A

HALT

The LD A, (addr) and LD (addrl,A instructions contain addresses to determine the source
or destination of the data. The addresses are 16 bits long, with the eight least Significant bits In the word Immediately following the instruction code and the eight most significant bits in the next word (this order IS contrary to normal computer practice). CPL IS
a one-word instruction that Inverts each bit of the Accumulator. It replaces each '0'
with a T and each T with a '0', Just like a set of Inverter gates.
HALT IS used to end all the examples.
Note that we could also place an address into Registers Hand L and then use that address throughout the program. This is shown in the following program.

4-3

Source Program:
LD
LD
CPL
INC
LD
HALT

HL,40H
A,(HL)

POINT TO OPERAND
GET DATA
COMPLEMENT
POINT TO DESTINATION
;STORE RESULT

HL
(HL),A

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007

21
40
00
7E
2F
23

Which version do

77
76
YOU

Instruction
(Mnemonic)
LD

HLAOH

LD
CPL
INC
LD
HALT

A,(HL)
HL
(HL),A

think is better?

The two versions require the same number of bytes of memory even though the second
version is two instructions longer, This is because the second version uses fewer explicit addresses,

8-Bit Addition
Purpose: Add the contents of memory locations 0040 and 0041, and place the result
Into memory location 0042.
Sample Problem:

Result:

(0040)
(0041)

38
2B

(0042)

63

Source Program:
LD
LD
LD
ADD
LD
HALT

A,(40H)
B,A
A,(41H)
A,B
(42Hl.A

:GET FIRST OPERAND

;SAVE FIRST OPERAND
:GET SECOND OPERAND
;ADD OPERANDS
;STORE SUM

4-4

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB

3A
40
00
47
3A
41
40
80
32
42
00
76

Instruction
(Mnemonic)
LD

A (40H)

LD
LD

B.A
A.(41H)

ADD
LD

AB
(42Hl.A

HALT

Here again. we could alternatively use Registers Hand L as the source for all addresses.
Source Program:

LD
LD
INC
ADD
INC
LD
HALT

HLAOH
A(HU
HL
A.(HU
HL
(HU.A

;GET FIRST OPERAND

;ADD SECOND OPERAND
;STORE RESULT

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008

21
40
00
7E
23
86
23
77

76

Instruction
(Mnemonic)
LD

HL,40H

LD
INC
ADD
INC
LD
HALT

A.(HU
HL
A.(HU
HL
(HU.A

In thiS case. the program using Registers Hand L IS Shorter than the one using direct
addressing. Why?
LD HL,40H loads the contents of the following two words of program memory Into
Register Pair HL. The first word goes into Register L. the second Into Register H.
The code (HU means that data IS obtained from or sent to the memory location addressed by Registers Hand L. Thus. LD A.(HU loads the Accumulator with the contents
of the addressed memory location; LD (HU.A loads the addressed memory location with
the contents of the Accumulator. ADD A(HU adds the contents of the location addressed by HL to the contents of the Accumulator. Remember that Hand L contain a
16-blt address. but the memory location with that address contains eight bits of data.
Note the difference between ADD A.(HU and ADD A.H or ADD A.L.
INC HL performs a 16-bit Increment in one instruction cycle. The CPU doesn't use the
8-bit arithmetic Unit for the Increment; It uses the incrementer that It normally uses to
increment the 16-bit Program Counter.

4-5

LDA.(HU and LD (HU,A are preferable to LD A.(addr) and LD (addr).A whenever you
use the same memory location repeatedly or use adjacent locations. because LD A(HU
and LD (HU.A require less program memory and time. Note. however. that you must
load Registers Hand L before you can use (HL).

Shift Left One Bit

Purpose: Shift the contents of memory location 0040 left one bit and place the result
into memory locatiOn 0041. Clear the empty bit position. This type of shift IS
known as a logical shift. In a logical shift. a value of zero is always shifted in.
Sample Problem:

Result:

(0040)

6F

(0041)

DE

Source Program:
LD
ADD
LD
HALT

A.(40H)
A.A
(41H).A

:GET DATA
:SHIFT LEFT
:STORE RESULT

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007

3A
40
00
87
32
41
00
76

Instruction
(Mnemonic)
LD

A.(40H)

ADD
LD

A.A
(41H).A

HALT

ADD A.A Simply adds the contents of the Accumulator to itself. The result. of course. is
twice the onginal data. which is the same result that a logical left shift would produce.
The least Significant bit of the result IS zero. since 0+0 = 1+1 = 0: 1+1 also produces a
Carry to the next bit.
Alternatively. we could replace ADD A.A with SLA A. certainly the more obvious
choice. However. SLA A requires two words of program memory and eight clock cycles.
while ADD A.A requires one word of program memory and four clock cycles. The
difference IS caused by the fact that SLA A is one of the extra instructions added to the
original 8080A set (re;"ember the companson presented earlier).

Mask Off Most Significant Four Bits

Purpose: Place the least Significant four bits of memory location 0040 into the least
Significant four bits of memory location 0041. Clear the most significant four
bits of memory location 0041.
Sample Problem:

Resu It:

(0040)

3D

(0041)

OD

4-6

Source Program:

LD
AND
LD
HALT

A.(40H)
00001111B
(41H).A

:GET DATA
:MASK 4 LSB'S
:STORE RESULT

Note: B means binary In standard Z80 assembler notation.

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008

3A
40
00
E6
OF
32
41
00
76

Instruction
(Mnemonic)
LD

A.(40H)

AND

00001111B

LD

(41H),A

HALT

The mask (00001111) IS written In binary to make ItS function clearer to the reader. Binary notation for masks is generally much clearer than hexadeCimal notation. although
the resu Its are the same. Hexadecimal notation shou Id be used for masks longer than
four bits. The comments should explain the masking operation.
When the argument In the address field IS a number. AND logically ANDs the contents
of the Accumulator with the contents of the word of program memory immediately
following the instruction. AND may be used to clear bits that are not in use. The four
least significant bits could be an Input from a switch or an output to a numeric display.

Clear a Memory Location

Purpose: Clear memory location 0040.
Source Program:

SUB
LD
HALT

A
(40H),A

:CLEAR LOCATION 40

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004

97

Instruction
(Mnemonic)
SUB
LD

32
40
00
76

A
(40H).A

HALT

SUB A subtracts the number in the Accumulator from Itself. The result IS to clear the
Accumulator. SUB A. XOR A. or LD A.O can all clear the Accumulator. LD A.O takes
more time and memory but doesn't affect the status flags.

Word Disassembly
Purpose: Divide the contents of memory location 0040 into two 4-blt sections and
store them In memory locations 0041 and 0042. Place the four most significant bits of memory location 0040 into the four least Significant bit positions

4-7

of memory location 0041, place the four least significant bits of memory
location 0040 into the fou r least significant bit positions of memory location
0042. Clear the four most significant bit positions of memory locations 0041
and 0042.

Sample Problem:

Result:

(0040)

3F

(0041)
(0042)

03
OF

Source Program:
LD
LD
LD
RRA
RRA
RRA
RRA
AND
INC
LD
LD
AND
INC
LD
HALT

HLAOH
A.(HU
B,A

:GET DATA
;SHIFT DATA RIGHT 4 TIMES

00001111B
HL
(HLl,A
A,B
00001111B
HL
(HL),A

:MASK OFF MSB'S

;STORE MSB'S
:RESTORE ORIGINAL DATA
;MASK OFF LSB'S
:STORE LSB'S

Object Program:
Memory Address
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
OOOB
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012

21
40
00
7E
47
1F
1F
lF
1F
E6
OF
23

77
78
E6
OF
23

77
76

Instruction
(Mnemonic)
LD

HLAOH

LD
LD
RRA
RRA
RRA
RRA
AND

A.(HL)
B,A

00001111B

INC
LD
LD
AND

HL
(HLl,A
A,B
00001111B

INC
LD
HALT

HL
(HLl,A

Instructions usmg the address in Registers Hand L occupy only one word of program
memory. However, HL must be loaded before the address can be used. Thus, implied
memory addressing saves time and memory, as compared to direct memory addressing,
only when the program repeatedly uses the same address or consecutive addresses.

4-8

RRC shifts the Accumulator right one bit circular. with the least significant bit going to
the most significant bit positiOn and to the Carry. Shifting the Accumulator right four
times requires four RRCs. We could use SRL A to provide a logical shift directly (no final
AND would then be necessary). However. SRL A requires tWice as much time and
memory as RRC. Try substituting SRL A for RRC and see the difference. Another alternative would be to use the RLD instruction to replace both the mask and the store.
However. this solution IS not optimal In terms of either storage or execution speed due
to the constraint that the high-order nibble of each result must equal zero.
Many Z80 instructions affect a pair of 8-bit registers. The pairs are HL (H and L). DE (0
and El. and 8C (8 and Cl. Registers B. D. and H are the most significant eight bits of the
pairs; Registers C. E. and L are the least significant eight bits. The common instructions
that use pairs of registers are LD rp (Load Register Pair). INC rp (Increment Register
Pair). DEC rp (Decrement Register Pair). and .ADD HL.rp (Add Register Pair to Hand L).

Find Larger of Two Numbers

Purpose: Place the larger of the contents of memory locations 0040 and 0041 Into
memorv location 0042. Assume that the contents of memorv locations 0040
and 0041 are unsigned binary numbers.
Sample Problems:

a
Result:
b
Result:

(0040)
(0041)

3F
2B

(0042)

3F

(0040)
(0041)

75
A8

(0042)

A8

Source Program:

DONE.

LD
LD
INC
CP
JR
LD
INC
LD
HALT

HL.40H
A.(HL)
HL
(HL)
C.DONE
A.(HL)
HL
(HL).A

:GET FIRST OPERAND

:IS SECOND OPERAND LARGER?
;YES. GET SECOND OPERAND INSTEAD
;STORE LARGER OPERAND

Object Program:

Memory Address
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB

21
40
00
7E
23
BE
30
01
7E
23
77
76

Instruction
(Mnemonlc)

DONE:

4-9

LD

HLAOH

LD
INC
CP
JR

A(HL)
HL
(HL)
NC.DONE

LD
INC
LD
HALT

A(HL)
HL
(HL).A

CP (HL) sets the flags as if the contents of the memory location addressed by Hand L
had been subtracted from the contents of the Accumulator. However, the contents of
the Accumulator are left unchanged for later comparisons or other processing.
If A is the contents of-the Accumulator and X is the second operand for a CP instruction, then the flags are set as follows:
1)
2)

Zero = 1 if A = X
Zero = 0 if A f:. X
Carry = 1 if A < X
Carry = 0 if A 2;; X
(A, X are unsigned binary numbers)

CP sets the Carry to 1 if a borrow would be necessary to actually perform the subtraction. I.e.. if the number being subtracted from the contents of the Accumulator is
greater than those contents. Thus, the sequence CPo JR NC.DONE causes a Jump to
DONE if the contents of the Accumulator are greater than or equal to the other number.
JR NC,DONE causes a jump to memory location DONE if the Carry flag = O. Otherwise
(if Carry = 1). the computer continues with the next sequential memory location after
the JR instruction.
DONE is a label. a name which you assign to a location In memory so that it IS easier to
remember. Note that labels are followed by a colon on the line where they are defined.
The label makes the destination of the branch clearer. particularly when relative addressing IS being used. The assembler calculates the required offset (caution: some zao
assemblers will not do thiS). Using a label is preferable to Just specifying the offset (j.e..
JR NC.$+3) since the zao's instructions vary in length. You could therefore easily make
an error In determining an offset.
If the branch conditions are not satisfied, the processor simply proceeds to the next sequential location In program memory (i.e.. it executes the instruction LD A,(HL}l.
The zao assemblers allow SIX characters in labels - the first must be a letter. while the
others may be letters or numbers (some special characters are allowed but we will not
use them).
The JR instruction uses relative addressing in which the second word of the Instruction
IS an 8-blt twos complement number that the CPU adds to the address of the next instruction to find the target address. In the example, the relative offset is 0009 (target
address) minus 0008 (address Immediately follOWing the branch) or 01.
We should note that some Z80 assemblers will not calculate the offset In the form
shown. These assemblers require an offset in the address field, rather than the label of
the target instruction. If you have such an assembler. use the form JR NC,DONE-$.
Remember that $ means "the address of the current instruction"
The Z80 has two sets of Jump Instructions, JP (Jump) and JR (Jump RelatiVe). The JP
Instructions require a complete memory address; they occupy three bytes of memory
and execute In ten clock cycles. The JR instructions require only a one-word offset:
they occupy two bytes of memory and execute in 12 cycles if a jUmp is actually performed and in 7 if not. So the JR instructions use less memory than JP instructions but
may require a little extra time if a jump IS performed hhe extra time IS used to execute
the required 16-blt addition of program counter and offset).

4-10

16-Bit Addition
Purpose: Add the 16-bit number in memory locations 0040 and 0041 to the 16-bit
number in memory locations 0042 and 0043. The most significant eight bits
are in memory locations 0041 and 0043. Store the result in memory locations 0044 and 0045. with the most significant bits in 0045.
Sample Problem:

(0040)
(0041)
(0042)
(0043)

2A
67
F8
14

Result: 672A + 14F8 = 7C22

(0044)
22
(0045) = 7C
Source Program:

LD
LD
ADD
LD
HALT

HL.(40H)
DE.(42H)
HL.DE
(44H),HL

:GET FIRST 16-81T NUM8ER

:GET SECOND 16-BIT NUMBER
.16-BIT ADDITION
:STORE 16-BIT RESULT

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB

2A
40
00
ED
58
42
00
19
22

Instruction
(Mnemonic)
LD

HL.(40H)

LD

DE.(42H)

ADD
LD

HL.DE
(44H).HL

44
00
76

HALT

LD HL,(addr) loads Registers Hand L from two memory locations. the one specified in
the instruction and the next consecutive one. The contents of the first addressed location go to Register L. The contents of the next location go to Register H. Thus. LD
HL,(40H) means L = (40), H = (41), The actual transfer proceeds one byte at a time and
takes 16 clock cycles. The advantage of the 16-bit Load Instruction over two 8-bit Load
Instructions is that the CPU has to fetch only one instruction from memory.
Note the difference betweell LD HL,(addr). which loads the contents of the two RAM
locatIOns at addr and addr+ 1 into Hand L. and LD HL.data16. which loads the contents
of the next two bytes pOinted to by the instruction counter into Hand L. Since these
two bytes immediately follow the op-code. loads of thiS type are referred to as load immediate instructions.
LD DE.(addr) is similar to LD HL,(addr) except that it takes one extra word of memory
and four more clock cycles. This is one of the Instructions that is present in the Z80 set
but not in the 8080/8085 sets. An alternative approach is:
EX
LD

DE.HL
HL,(42H)

:SAVE FIRST 16-BIT NUMBER IN DE

:GET SECOND 16-BIT NUMBER
4-11

EX DE.HL exchanges the contents of Registers D and E with Hand L. No numbers are
changed or destroyed. The advantage of EX DE.HL will become obvIous if you try to
replace it with a series of LD instructions.
ADD HL.DE adds the 16-blt number In Registers D and E to the 16-bit number in
Registers Hand L. The result IS placed into Registers Hand L. ADD HL.DE actually adds
one byte at a time. It executes in 11 clock cycles.
LD (addrl.HL stores the contents of Registers Hand L Into two memory locatIOns. the
one specified In the instruction and the next consecutive one. The contents of L go Into
the specified location and the contents of H go Into the next location. Thus. LD (44Hl.HL
means (44) = L. (45) = H.As with LD HL.(addrl. the actual transfer proceeds one byte at
a time and requires 16 clock cycles.
Although the zao IS an a-bit processor. It has instructions that handle 16-bit numbers.
These instructions are Intended pnmarily for handling addresses. but you can also use
them for 16-blt data. The most common ones and their uses are:
1)
2)
3)
4)

ADD HL,rp -16-Bit Add

Used to access tables and to add 16-blt data units
DEC rp - 16-8it Decrement
Used to subtract one from the contents of a register pair
INC rp - 16-8it Increment
Used to add one to the contents of a register pair
LD rp,data16 -16-8it Load Immediate
Used to Initialize a register pair with a fixed value, e.g.. the starting address of an array or table

5)

LD HL.(addrl- 16-8it Load HL Direct

Used to place vanable addresses Into the main address register (H and L)

6)

LD (addrl.HL - 16-8it Store HL Direct

Used to store addresses to memory from the main address register (H and L).

Table of Squares
Purpose: Calculate the square of the contents of memory location 0040 from a table
and place it Into memory location 0041. Assume that memory location 0040
contains a number between 0 and 7 inclusive (0 ~ (0041) ~ 7).

The table occupies memory locations 0050 to 0057

Entry

Memory Address
(Hex)

(Hex)

0050
0051
0052
0053
0054
0055
0056
0057

00
01
04
09
10
19
24
31

(Decimall
0
1
4
9
16
25
36
49

Sample Problems:

a.

(0041)
Result:

b.
Result:

03

(0042)

09

(0041)

06

(0042)

24

4-12

(0 2)
(12)
(22)
(3 2)
(4 2)
(52)
(6 2)
(7 2)

Source Program:
LD
LD
LD
LD
ADD
LD
LD
HALT
SQTAB:

ORG
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB

A.(40H)

:GET DATA
;MAKE DATA INTO 16-BIT INDEX

L.A

H.O
DE,SOTAB
HL.DE
A,(HL)
(41H).A

;GET STARTING ADDRESS OF TABLE

;INDEX TABLE WITH DATA
;GET SQUARE OF DATA

50H

;SQUARE TABLE

o
1
4
9

16
25
36

49

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009

3A
40
00

LD

A.(40H)

6F

LD
LD

L.A

LD

DE.SOTAB

ADD
LD
LD

HL.DE
A.(HL)
(41H).A

OOOA
OOOB

oooe

0000

OOOE
0050
0051
0052
0053
0054
0055
0056
0057

Instruction
(Mnemonic)

26
00
11

H.O

50
00
19

7E

32
41

00
76
00
01
04
09
10

SQTAB:

19

24
31

DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB

0
1
4
9
16

25
36
49

Note that you must also enter the table of squares into memory (the assembler pseudooperation DEFB will handle this!' The table of squares is constant data, not parameters
that may change; that is why you can initialize the table using the DEFB pseudo-operation, rather than by executing instructions to load values into the table. Remember that
the table is part of the program memory (ROM in most systems).
LD LA moves the data in the Accumulator to Register L. The data IS the eight least significant bits of the index. You cannot always assume that the data presented to your

4-13

program is in the proper range. It is always a good practice to range check all critical
values. Range checking consists of testing a value to ensure that It is within the proper
lower and upper limits. Any byte can have a value in the range 0 to 255. If the value
stored in the byte at location 0040H is greater than seven, the program will reference
an undefined byte beyond the end of the square table, causing the program to generate
erroneous results Range checking will prevent this error from occurring.
LO H,O clears Register H so that it does not interfere with the 16-bit addition of starting
address and index. Never assume that a register contains zero at the start of a program.
LO OE.SQTAB loads the starting address of the table Into Registers 0 and E. We use 0
and E for the starting address since the ADD HL Instruction does not change 0 and E.
Thus, the starting address of the table will still be in 0 and E after the addition, in the
event that we want another element from the table.
ADD HL,OE adds the starting address and the Index: the result in Hand L is thus the address of the correct entry. LO A.(HL) then moves that entry to the Accumulator.
Arithmetic that a microprocessor cannot do directly in a few instructions is often best
performed with lookup tables. Lookup tables Simply contain all the possible answers to
the problem: they are organized so that the answer to a particu lar problem can be
found easily, The arithmetic problem now becomes an accessing problem - how do
we get the correct answer from the table? We must know two things: the position of
the answer in the table (called the index) and the base. or starting, address of the table.
The address of the answer is then the base address plus the index,
The base address. of course. is a fixed number for a particular table, How can we determine the index? In simple cases. where a single piece of data is involved. we can organize the table so that the data is the Index. In the table of squares, the Oth entry in the table contains zero squared. the first entry one squared, etc. In more complex cases,
where the spread of input values is very large or there are several data items involved
(e.g.. roots of a quadratic or number of permutations), we must use more. complicated
methods to determine indexes.
The basic tradeoff in using a table is time vs. memory. Tables are faster. since no computations are required. and simpler. since no mathematical methods must be devised
and tested. However, tables can occupy a large amount of memory if the range of the
input data is large, We can often reduce the size of a table by limiting the accuracy of
the results, scaling the input data. or organizing the table cleverly. Tables are often
used to compute transcendental and trigonometric functions. linearize inputs. convert
codes, and perform other mathematical tasks.

16-Bit Ones Complement

Purpose: Place the ones complement of the 16-bit number in memory locations 0040
and 0041 into memory locations 0042 and 0043, The most significant bytes
are in locations 0041 and 0043.
Sample Problem:

Result:

(0040)
(0041)

67
E2

(0042)
(0043)

98
10

The ones complement inverts each bit of the original number: the sum of the original
number and its ones complement will always be all 1 bits.

4-14

Source Program:

LD
LD
CPL
LD
LD
CPL

HL,{40Hl
A.L

A.H

:COMPLEMENT 8 MSS'S

LD

H.A
(40H).HL

:STORE ONES COMPLEMENT

LD
HALT

:GET DATA
:COMPLEMENT 8 LSS'S

LA

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOS
OOOC

2A
40
00
7D
2F
6F
7C
2F
67
22
42
00
76

Instruction
(Mnemonic)
LD

HL,{40H)

LD
CPL
LD

A.L

LD

LA
A.H

CPL
LD
LD

H.A
(42H).HL

HALT

Despite the Z80' s 16-bit instructions. you must use 8-bit instructions to perform most
arithmetic and logical operations. The 16-bit instructions can. however. be used to load
and store data and occasionally to do a few 16-bit arithmetic operations. such as addition. subtraction. incrementing. and decrementing. You will soon learn that the 16-bit
instructions are far from a complete set and you may often run into awkward problems
if using them to manrpulate 16-blt data.

PROBLEMS
1) Twos Complement
Purpose: Place the twos complement of the contents of memory location 0040 into
memory location 0041. The twos complement is the ones complement plus
one.
Sample Problem:

Result:

(0040)

3E

(0041)

C2

The sum of the original number and its twos complement is zero (try the sample case).

2)

8-Bit Subtraction

Purpose: Subtract the contents of memory location 0041 from the contents of memory
location 0040. Place the result into memory location 0042.
Sample Problem:

Result:

(0040)
(0041)

77
39

(0042)

3E

4-15

~--~ ..

_---

3)

Shift Left Two Bits

Purpose: Shift the contents of memory location 0040 left two bits and place the result

into memory location 0041. Clear the two least significant bit positions.
Sample Problem:

Result:

4)

(0040)

50

(0041)

74

Mask Off least Significant Four Bits

Purpose: Place the four most significant bits of the contents of memory location 0040

into memory location 0041. Clear the four least significant bits of memory
location 0041.
Sample Problem:

Resu It:

5)

(0040)

C4

(0041)

CO

Set a Memory location to All Ones

Purpose: Memory location 0040 IS set to all ones (FF hex).

6)

Word Assembly

Purpose: Combine the four least significant bits of memory locations 0040 and 0041

Into a word and store them in memory location 0042. Place the four least significant bits of memory location 0040 Into the four most significant bit POSItions of memory location 0042; place the four least Significant bits of memory location 0041 Into the four least Significant bit positions of memory location 0042.
Sample Problem:

Result:

7)

(0040)
(0041)

6A
83

(0042)

A3

Find Smaller of Two Numbers

Purpose: Place the smaller of the contents of memory locations 0040 and 0041 Into

memory location 0042. Assume that 0040 and 0041 contain unsigned binary numbers.
Sample Problems:

a.
Result:
b.
Result:

8)

(0040)
(0041)

3F
2B

(0042)

28

(0040)
(0041)

75
A8

(0042)

75

24-Bit Addition

Purpose: Add the 24-blt number in memory locations 0040.0041. and 0042 to the 24bit number in memory locations 0043. 0044. and 0045. The most significant
eight bits are in memory locations 0042 and 0045: the least significant eight
bits are in memory locations 0040 and 0043. Store the result In memory
locations 0046. 0047. and 0048 with the most significant bits in 0048 and
the least significant bits in 0046.

4-16

Sample Problem:

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)

2A
67
35
F8
A4
51

(0046)
(0047)
(0048)

22
OC
87

Result:

that is, 35672A

+51A4F8
870C22

9)

Sum of Squares

Purpose: Calculate the squares of the contents of memory locations 0040 and 0041
and add them together, Place the result Into memory location 0042. Assume
that memory locations 0040 and 0041 both contain numbers between 0 and
7 inclusive (0 ~ (0040) ~ 7 and 0 ~ (0041) ~ 7), Use the table of squares
from the example entitled Table of Squares.
Sample Problem:

(0040)
(0041)

03
06

(0042)
2D
2
2
that IS, 3 +6 = 9 + 36 =45
2D (hex)

Result:

10)

16-Bit Twos Complement

Purpose: Place the twos complement of the 16-blt number In memory locations 0040
and 0041 (most Significant bits in 0041) Into memory locations 0042 and
0043 (most significant bits In 0043).
Sample Problems:

a.
Result:
b.
Result:

(0040)

00

(0041)

58

(0042)
(0043)

00
A8

(0040)
(0041)

72
00

(0042)
(0043)

8E
FF

4-17

Chapter 5
SIMPLE PROGRAM LOOPS
The program loop IS the basIc structure that forces the CPU to repeat a sequence of instructions. Loops have four sections:
1)

The initialization section. which establishes the starting values of counters. address
registers (pointers), and other variables.

2)

The processing section. where the actual data manipulation occurs. This is the section that does the work.

3)

The loop control section. which updates counters and pointers for the next Iteration.

4)

The concluding section. which analyzes and stores the results.

Note that the computer performs Sections 1 and 4 once. while it may perform Sections
2 and 3 many times. Thus. the execution time of the loop will mainly depend on the execution time of Sections 2 and 3. You will want Sections 2 and 3 to execute as quickly
as possible; do not worry about the execution time of Sections 1 and 4. A typical program loop can be flowcharted as shown in Figure 5-1. or the positions of the processing
and loop control sections may be reversed as shown in Figure 5-2. The processing section in Figure 5-1 IS always executed at least once. while the processing section In
Figure 5-2 may not be executed at all. Figure 5-1 seems more natural. but Figure 5-2 IS
often more efficient and avoids the problem of what to do when there IS no data (a
bugaboo for computers. and the frequent cause of silly situations like the computer
dunning someone for a bill of $0.00).
The loop structure can be used to process entire blocks of data. To accomplish this. the
program must increment an address register (usually register pair HL) after each iteration so that the address register points to the next element in the data block. The next
iteration will then perform the same operations on the data in the next memory location. The computer can handle blocks of any length with the same set of instructions.
Implied addressing through register pairs (particularly HL) is the key to processing a
block of data with the
since it allows you to vary the actual memory address by
changing the contents of registers. Indexed addressing. while longer and slower on the
than implied addressing, may be handy when processing more than one block of
data. Note that In the immediate and direct addressing modes. the addresses that are
used are completely determined by the instruction (and thus fixed if the program
memory IS read-only).
.

zao.

zao

5-1

Initialization
SectIon

Processing
SectIon

Loop Control
Section

Concluding
Section

Figure 5-1. Flowchart of a Program Loop

5-2

Initialization
Section

Ves

Processing

Concluding

Section

Section

Figure 5-2. A Program Loop that Allows Zero Iterations

EXAMPLES
Sum of Data
Purpose: Calculate the sum of a series of numbers. The length of
the series IS In memory location 0041. and the series
begins In memory location 0042. Store the sum in
memory location 0040. Assume that the sum IS an 8-bit
number so that you can Ignore carnes.
Sample Problem:

Result:

(0041)
(0042)
(0043)
(0044)

03
28
55
26

(0040)

(0042) + (0043)
28+55+26

(0044)

A3
There are three entries

In

the sum, since (0041 )=03.

5-3

8BIT
SUMMATION

Flowchart:

POint.r
Count

=
=

41
(Point.rl

Sum = 0

POInter :; Pointer + 1
Sum = Sum

+ (Polnt.rl

Count

(401

Count - 1

Sum

Note: (Pointed is the contents of the memory location addressed by Pointer. Remember
that on the ZSO, Pointer IS a 16-bit address, while (Pointed is an S-blt byte of
data.
Source Program:

SUMO:

LD
LD
SUB
INC
ADD
DEC

JR
LD
HALT

HL.41 H
B,(HL)

:COUNT LENGTH OF SERIES

:SUM = ZERO

HL
A,(HL)

:SUM = SUM

B
NZ,SUMD
(40H).A

:STORE SUM

5-4

+ DATA

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000

21
41
00
46
97
23
86
05
20
FB
32
40
00
76

Instruction
(Mnemonic)

SUMO:

LD

HL.41 H

LD
SUB
INC
ADD
DEC
JR

B.(HU
A
HL
A.(HU
B
NZ.SUMD

LD

(40H),A

HALT

The initialization section of the program IS the first three instructions which set the sum.
counter. and data pointer to their starting values.
Note that you can use LD to transfer data between memory and any of the primary
general purpose registers O.e.. A. B. C. D. E. H. U uSing the address in Registers Hand L.
However. the only transfers allowed uSing direct addreSSing are those that move data to
or from the Accumulator O.e.. LD A.(addrl and LD (addrl.A - there IS no instruction LD
E.(addr). for example),
The processing section of the program is the single Instruction ADD A.(HU which adds
the contents of the memory location being addressed by Registers Hand L to the contents of the Accumulator. and stores the result In the Accumulator. This instruction
does the real work of the program.
The loop control section of the program consists of the instructions INC HL and DEC B.
INC HL updates the pOinter so that the next Iteration adds the next number to the sum.
DEC B decrements the counter that keeps track of how many iterations are left.
The Instruction JR NZ causes a branch if the Zero flag is zero. The offset IS a twos complement number. and the count begins from the memory location Immediately followIng the JR instruction. In this case. the reqUired lump is from memory location OOOA to
memory location 0005. So the offset IS:
0005
-OOOA

05
+F6

FB

If the Zero flag IS one. the CPU executes the next instruction In sequence O.e.. LD
(40H),A), Since DEC B was the last instruction before JR to affect the Zero flag. JR
NZ.SUMD causes a Jump to SUMO if DEC B does not produce a zero result. I.e..
SUMO if B #=0
PC =
{ PC+2 if B =0
(The 2 IS caused by the two-word JR Instruction!'

5-5

The loop control sequence DEC followed by JR NZ IS so common that the Z80 has a
special instruction that both decrements the counter and performs the jump. This instruction IS DJNZ. Decrement and Jump on Not Zero, which decrements Register Band
then lumps by the specified relative offset if the remainder IS not zero. So we could
change the end of the example to:
DJNZ
LD
HALT

SUMO
(40Hl.A

Which has the object form:

10

07
08
09
OA
OB
OC

DJNZ

SUMO

LD

(40Hl.A

FC
32
40
00
76

HALT

This change saves one byte of memory and three clock cycles. Note, however, that you
must use Register B as the counter since this is the register that DJNZ decrements.
Since the offset in Z80 relative lumps IS only one byte long. such lumps can go no
further than 127 locations forward or 128 locations backward (actually 129 forward or
126 backward, since the count starts at the end of the 2-word instruction). Longer
jumps must use the JP instructions.
Most computer loops count down rather than up so that the Zero flag can serve as an
eXit condition. Remember that the Zero flag IS 1 if the result was zero and 0 if the result
was not zero. Try rewriting the program so that it counts up rather than down: which
method IS more effiCient?
The order of instructions is often very Important. DEC B must come right before JR
NZ,SUMD, since otherwise the Zero result set by DEC B could be changed by another
instruction. INC HL must come before ADD A. (HL) or else the first number added to the
sum will be the contents of memory location 0041 Instead of the contents of memory
location 0042.

16-Bit Sum of Data

Purpose: Calculate the sum of a series of numbers. The length of the series IS In
memory location 0042 and the series itself beginS In memory location 0043.
Store the sum in memory locations 0040 and 0041 (eight least significant
bits In 0040).
Sample Problem:

(0042)
(0043)
(0044)
(0045)

03
C8
FA
96

Result: C8 + FA + 96 = 0258
(0040)
58
(0041) = 02

5-6

Flowchart:

POinter
Count

:=

42
(Pointer!

Suml = 0
Sumu = 0

Pointer "" Pointer + 1

Suml =Suml

+ (PolOterl

Sumu

Count

-=::

Sumu + 1

Count - 1

No

(40) = Suml
(411 ~ Sumu

Source Program:
LD
LD
SUS
LD
DSUMD: INC
ADD
JR
INC
CHCNT DJNZ
LD
LD
INC
LD
HALT

HLA2H
S,(HU

A
C,A
HL
A,(HU
NC,CHCNT

:COUNT = LENGTH OF SERIES

:LSS'S OF SUM = 0
:MSB'S OF SUM = 0
:SUM = SUM

+ DATA

:ADD CARRY TO MSS'S OF SUM

DSUMD
HLAOH
(HU,A
HL
(HU,C

:STORE LSS'S OF SUM

:STORE MSS'S OF SUM

5-7

Object Program:
Memorv Address
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012
0013

21
42
00
46
97
4F
23
86
30
01
OC
10
F9
21
40
00

Instruction
(Mnemonic)
LD

LD
SUB
LD
DSUMD: INC
ADD
JR

CHCNT'

77
23
71
76

HLA2H

B.(HL)
A
C.A
HL
A.(HL)
NC.CHCNT

INC
DJNZ

C
DSUMD

LD

HL40H

LD
INC
LD
HALT

(HL).A
HL
(HL).C

The structure of this program IS the same as the structure of the last one. The most significant bits of the sum now must be initialized and stored. The processing section consists of three instructions (ADD A.(HL); JR NC.CHCNT; and INC Cl. Including a Conditional Jump.
JR NC.CHCNT causes a lump to memory location CHCNT if the Carry = O. Thus. if there
IS no carry from the 8-blt addition. the program jumps around the statement that increments the most significant bits of the sum. The relative offset IS:
0008
-COOA

--0-1
The relative offset for DJNZ DSUMD

IS:

0006

06

-OOOD

+F3

F9
INC C adds 1 to the contents of Register C. Note that INC BC IS a 16-blt Increment that
adds 1 to Register C and adds the resulting carry to Register B; INC C IS an 8-bit increment that does not account for the carry.

5-8

Number of Negative Elements

Purpose: Determine the number of negative elements (most significant bit 1) in a
block. The length of the block is in memory location 0041 and the block Itself
starts in memory location 0042. Place the number of negative elements in
memory location 0040.
Sample Problem:

(0041)
(0042)
(0043)
(0044)
(0045)
(0046)
(0047)
Result:

(0040)

06

68
F2

87
30

59

2A
02. since 0043 and 0044 contain
numbers with an MSB of 1.

Flowchart:

Pomter = 41
Count = (Pointerl
Nne9 = 0

Nne9 = Nne9 + 1

Count = Count - 1

No

(401

= Nne9

5-9

Source Program:

SRNEG:

CHCNT

LD
LD
LD
INC
LD
AND
JP
INC
DJNZ
LD
LD
HALT

HL,41H
B.(HL)
C.O
HL
A.(HU

:COUNT = NUMBER OF ELEMENTS

:NUMBER OF NEGATIVES = ZERO
:GET NEXT ELEMENT
;JS MSB ZERO?

A
P.CHCNT

:NO. ADD 1 TO NUMBER OF NEGATIVES

SRNEG
A.C
(40H),A

:STORE NUMBER OF NEGATIVES

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
OOOB
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012
0013

21
41
00
46
OE
00
23
7E
A7
F2
OD
00
OC
10
F7
79
32
40
00
76

Instruction
(Mnemonic)

SRNEG:

CHCNT:

LD

HL,41H

LD
LD

B.(HU
C.O

INC
LD
AND
JP

HL
A.(HU
A
P.CHCNT

INC
DJNZ

C
SRNEG

LD
LD

A.C
(40H),A

HALT

AND A simply sets the flag bits according to the contents of the Accumulator without
affecting those contents: OR A has the same effect. This is necessary since merely loadIng the Accu mu latar does not affect the flags.
JP P.CHCNT requires a full 16-blt address. There
there IS on the Carry and Zero flags.

IS

no relative Jump on the Sign flag like

Note that all we really want to do IS test the value of bit 7 of the memory location addressed by Registers Hand L. The zao has a special bit testing instruction. BIT. that IS
designed specifically for thiS purpose. BIT sets the Z flag to the complement of the indicated bit Within the Indicated register or memory location. For example. BIT 5.D will set
Z to 1 if bit 5 of Register D IS zero. and to 0 if bit 5 of Register D IS one. An implementation of this alternative IS as follows.

5-10

Source Program:

SRNEG:

CHCNT

LD
LD
LD
INC
BIT
JR
INC
DJNZ
LD
LD
HALT

HL,41 H
B.(HL)
C.O
HL
7.(HL)
Z.CHCNT

;COUNT = NUMBER OF ELEMENTS

:NUMBER OF NEGATIVES = ZERO
:IS NEXT ELEMENT NEGATIVE?

;YES. ADD 1 TO NUMBER OF NEGATIVES

SRNEG
A.C
(40Hl.A

:STORE NUMBER OF NEGATIVES

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012

21
41
00
46
OE
00
23
CB
7E
28
01
OC
10
F8
79
32
40
00
76

Instruction
(Mnemonic)

SRNEG:

CHCNT

LD

HL,41 H

LD
LD

B.(HL)
C.O

INC
BIT

HL
7.(HL)

JR

Z.CHCNT

INC
DJNZ

C
SRNEG

LD
LD

A.C
(40Hl.A

HALT

BIT 7. (HU sets the Z bit if bit 7 of the memory location addressed by Registers Hand L IS
zero. and clears the Z bit if bit 7 of that location is one. BIT does not affect any registers
or memory locations.
This program uses JR Z.CHCNT since no incrementing
IS zero.

IS

necessary if the addressed bit

Still another approach would be to use the instruction RLC (HL) to shift the sign bit of
the data In memory to the Carry. The required jump would then be JR NC.CHCNT
However. this approach uses extra time (RLC (HL) takes 15 cycles as compared to the
12 needed by BIT 7. (HL)) and also changes the data In memory which may be needed
for other purposes. Note that these disadvantages are related: the extra time IS needed
to return the result to the memory location.

Find Maximum
Purpose: Find the largest element in a block of data. The length of the block IS In
memory location 0041 and the block itself begins in memory location 0042.
Store the maximum In memory location 0040. Assume that the numbers in
the block are all 8-blt unsigned binary numbers.

5-11

Sample Problem:

(0041)
(0042)
(0043)
(0044)
(0045)
(0046)
Result:

(0040)

05

67
79
15
E3
72

E3, since this is the largest of

the five unsigned numbers,

Flowchart:

Pointer =
Count =

Max

Pointer

Max

41
(Pointer)

= 0

= POinter

+ 1

= (Pointer}

Count = Count - 1

No

1401 = Max

5-12

Source Program:

NEXTE:

DECNT:

LD
LD
SUB
INC
CP
JR
LD
DJNZ
LD
HALT

:POINT TO COUNT
:COUNT = NUMBER OF ELEMENTS
:MAXIMUM = MINIMUM POSSIBLE VALUE (ZERO)

HLA1H
B.(HU

A
HL
(HU
NC.DECNT
A.(HU
NEXTE
(40H).A

:IS NEXT ELEMENT ABOVE MAXIMUM?

:YES. REPLACE MAXIMUM WITH ELEMENT
:SAVE MAXIMUM

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF

21
41
00
46
97
23
BE
30
01
7E
10
F9
32
40
00
76

Instruction
(Mnemonic)

NEXTE:

DECNT'

LD

HLA1H

LD
SUB
INC
CP
JR

B.(HU
A
HL
(HU
NC.DECNT

LD
DJNZ

A.(HU
NEXTE

LD

(40H).A

HALT

The relatiYe offset for JR NC.DECNT IS:

OOOA
-0009
--0-1
The relative offset for DJNZ NEXTE IS:
0005
-OOOC

05
+F4

F9
The first three instructions of this program form the initialization section.
This program takes advantage of the fact that zero IS the smallest 8-blt unsigned binary
number. When you set the register that contains the maximum value - I n this case the
Accumulator - to the minimum possible value before you enter the loop. then the program will set the Accumulator to a larger value unless all the elements In the array are
zeros.

5-13

The program works properly if there are two elements. but not if there are one or none
at aU Why? How could you solve this problem?
The instruction CP (HU sets the Carry flag as follows (ELEMENT IS the contents of the
address In Registers Hand L and MAX IS the contents of the Accumulatod:
CARRY = 1 if ELEMENT> MAX
CARRY = 0 if ELEMENT ~ MAX
If CARRY = O. the program proceeds to DECNT and does not change the maximum. If
CARRY = 1. the program replaces the old maximum with the current element by executing the instruction LD A.(HU.
The program does not work if the numbers are signed because negative numbers will
appear to be larger than positive numbers. The problem IS somewhat tncky because
overflow could make the result appear to have the wrong sign.
Remember that overflow occurs when the magnitude of a result affects its sign bit. The
Z80 has a Panty/Overflow flag that indicates when twos complement overflow has occurred. Anthmetic operations that result In overflow set this flag. You can then test ItS
value With the instructIOns JP PE.ADDR (Jump on Panty Even - or Jump on Overflow)
or JP PO.ADDR (Jump on Panty Odd - or Jump on No Overflow). One thing you may
have to watch IS that thiS Z80 usage IS inconSistent With the 8080A or 8085
microprocessors. which always use the P flag to indicate panty. The 8080A and 8085
microprocessors have no overflow Indicator.

Justify a Binary Fraction

Purpose: Shift the contents of memory location 0040 left until the most Significant bit
of the number IS 1. Store the result In memory location 0041 and the number
of left shifts required In memory location 0042. If the contents of memory
location 0040 are zero. clear both 0041 and 0042.
Note: The process IS iust like converting a number to a SCientific notation:"ib?example:
0.0057 = 5.7 x 10-3
Sample Problems:

a.

(0040)

22

Result:

(0041)
(0042)

88
02

(0040)

01

Result:

(0041)
(0042)

80
07

(0040)

CB

Result:

(0041)
(0042)

CB
00

(0040)

00

Result:

(0041)
(0042)

00
00

c.

d.

5-14

Flowchart:

Nshft
Numb

0
140)

Yes

Yes

Shift Numb
lelt 1 bit
Nshlt = Nsht + 1

1411
142)

Numb
Nshlt

Source Program:

CHKMS.

DONE:

LD
LD
LD
AND
JR
JP
INC
ADD
JP
INC
LD
INC
LD
HALT

B.O
HL.40H
A.(HU
A
Z.DONE
M.DONE
B
A.A
CHKMS
HL
(HU.A
HL
(HU,B

NUMBER OF SHIFTS=ZERO
:GET DATA
IS DATA ZEROt
YES. DONE
;DONE IF SIGN BIT IS ONE
;ADD 1 TO NUMBER OF SHIFTS
:SHIFT LEFT ONE BIT

:SAVE JUSTIFIED DATA

:SAVE NUMBER OF SHIFTS

5-15

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012
0013
0014
0015

06
00
21
40
00
7E
A7
28
08
FA
11
00
04
87
C3
09
00
23

Instruction
(Mnemonic)

CHKMS.

DONE:

77
23
70
76

LD

B.O

LD

HLAOH

LD
AND
JR

A.(HU
A
Z.DONE

JP

M.DONE

INC
ADD
JP

B
A.A
CHKMS

INC
LD
INC
LD
HALT

HL
(HU.A
HL
(HU.B

JP M.DONE causes a Jump to location DONE if the Sign bit IS 1. This condition may
mean that the last result was a negative number or may Just mean that ItS most significant bit was 1 - the computer supplies only the results: the programmer must prOVide
the interpretation.
ADD A.A adds the number in the Accumulator to itself. The program uses this instructIOn. rather than RLA or RLCA. because ADD A affects the Sign bit while RLA and RLCA
do not.
We could reorganize this program so as to eliminate an extraneous JP and use relative
rather than absolute lumps. One reorganized version would be:
LD
LD
LD
AND
JR
DEC
CHKMS: INC
RLA
JR
RRA
INC
DONE:
LD
INC
LD
HALT

B.O
HL40H
A.(HU
A
Z.DONE
B
B
NC.CHKMS
HL
(HU.A
HL
(HU.D

:NUMBER OF SHIFTS = ZERO

:GET DATA
:IS DATA ZERO?
;YES. DONE
:ADJUST NUMBER OF SHIFTS BACK ONE
:ADD 1 TO NUMBER OF SHIFTS
:SHIFT LEFT ONE BIT
:CONTINUE IF MSB NOT ONE
:ADJUST DATA BACK
:SAVE JUSTIFIED DATA
:SAVE NUMBER OF SHIFTS

Show that this version also works. What are its advantages and disadvantages as compared to the prevIous program?

5-16

PROBLEMS
1)

Checksum of Data

Purpose: Calculate the checksum of a series of numbers. The length of the series

IS in
memorv location 0041 and the series itself begins m memory location 0042.
Store the checksum in memory location 0040. The checksum is formed bv
Exclusive-ORing all the numbers in the series together.

Note: Such checksums are often used m paper tape and cassette svstems to ensure
that the data has been read correctly. The calculated checksum is compared to
the one stored with the data - if the two checksums do not agree. the system
will usually either indicate an error to the operator or automatically read the data
agam.
Sample Problem:

(0041)
(0042)
(0043)
(0044)
Result:

2)

(0040)

03
28
55
26
(0042) EB (0043) EB (0044)
28EB55EB26
00101000
EB01010101
011 11101
EB00100110
01011011
5B

Sum of 16-Bit Data

Purpose: Calculate the sum of a series of 16-bit numbers. The length of the series

IS in
memory location 0042 and the series itself begms m memory location 0043.
Store the sum in memory locations 0040 and 0041 (eight most significant
bits in 0041). Each 16-bit number occupies two memory locations. with the
eight most significant bits in the higher address. Assume that the sum can
be contained in 16 bits.

Sample Problem:

(0042)
(0043)
(0044)
(0045)
(0046)
(0047)
(0048)

03
F1
28
1A
30
89
4B

Result: 28F1 + 301A + 4B89 =A494

(0040)
94
(0041) = A4

5-17

3)

Number of Zero, Positive, and Negative Numbers

Purpose: Determine the number of zero. positive (most significant bit zero but entire
number not zerol. and negative (most significant bit 1) elements In a block.
The length of the block IS in memory location 0043 and the block Itself starts
In memory location 0044. Place the number of negative elements In memory
location 0040. the number of zero elements In memory location 0041. and
the number of positive elements in memory location 0042.
Sample Problem:

(0043)
(0044)
(0045)
(0046)
(0047)
(0048)
(0049)
Resu It: 2 negative.
(0040)
(0041)
(0042)

4)

06
68
F2
87
00
59
2A
1 zero. and 3 positive. so
02
01
03

Find Minimum

Purpose: Find the smallest element In a block of data. The length of the block is In
memory location 0041 and the block Itself begins In memory location 0042.
Store the minimum In memory location 0040. Assume that the numbers In
the block are 8-blt unsigned binary numbers.
Sample Problem:

Resu It:

5)

(0041)
(0042)
(0043)
(0044)
(0045)
(0046)

05
67
79
15

(0040)

15. since this IS the smallest of the

five unsigned numbers.

E3
72

Count 1 Bits

Purpose: Determine how many bits in memory location 0040 are one and place the
result In memory location 0041.
Sample Problem:

Result:

(0040)

38 =00111011

(0041)

05

5-18

Chapter 6
CHARACTER-CODED DATA
Microprocessors often handle character-coded data. Not only do keyboards,
teletypewriters, communications devices, displays, and computer terminals expect or
provide character-coded data: many instruments, test systems, and controllers also reqUire data in this form. The most commonly used code IS ASCII. Baudot and EBCDIC are
found less frequently. We will assume all of our character-coded data to be 7-blt ASCII
with the most significant bit zero (see Table 6-1).
In handling ASCII-coded data are:
HANDLING
DATA IN
The codes for the numbers and letters form ordered sub-seASCII
quences. The codes for the decimal numbers are hex 30
through 39, so that you can convert between decimal and
ASCII with a simple additive factor. The codes for the upper-case letters are hex 41
through 5A. so that you can do alphabetic ordering by sorting the data In increasing numerical order.

Some principles to remember

1)

2)

The computer draws no distinction between printing and non-printing characters.

This distinction is made only by I/O devices.

3)

An ASCII device will handle only ASCII data. To print a 7 on an ASCII printer. the
microprocessor must send hex 37 to the printer: hex 07 is the 'bell' character.
Similarly, the microprocessor will receive the character 9 from an ASCII keyboard
as hex 39: hex 09 IS the 'tab' character.

4)

Some ASCII devices do not use the full character set. For example, control characters and lower-case letters may be Ignored or printed as spaces or question marks.

5)

Some widely used ASCII characters are:

OA
OD
20

16
16

- line feed (LF)

- carnage return (CR)
- space

16
3F 16 - ? (question mark)
7F 16 - rubout or delete character
6)

Each ASCII character occupies seven bits. This allows a large character set but IS
wasteful when the data is limited to a small subset such as the decimal numbers.
An a-bit byte, for example. can hold only one ASCII-coded decimal digit. while I'
can hold two BCD-coded digits.

6-1

Table 6-1. Hex-ASCII Table

NUL
SOH
STX
ETX
EOT
ENQ
ACK
BEL
BS
HT
LF
VT
FF
CR
SO
SI

OLE
DCl
DC2
DC3
DC4
NAK
SYN
ETB
CAN
EM
SUB
ESC
FS
GS
RS
US

SP

0
1
2
3
4
5
6

P
Q
R
S
T
U
V
W
X
Y

a
b
c
d
e
f
g
h

Hex LSD

0
1
2
3

6
6
7
8
9
A
B
C
D
E
F

I
"

#
$
%
&

7
(

.
)

A
B
C
0
E
F
G
H

<
=

>
?

K
L
M
N

Z
[

1
A

p
q
r
s
t
u

v
w
x

I
k
I
m

n
0

DEL

EXAMPLES
Length of a String of Characters
Purpose: Determine the length of a string of ASCII characters (seven bils with most
significant bit zerol. The string starts In memory location 0041. the end of
the string IS marked by a carriage return character ('CR', hex 001. Place the
length of the string (excluding the carriage returnl Into memory location
0040.
Sample Problems:
a.
Result:
b

Result:

(0041)

00

(0040)

00

since the first character is a camage return.

(00411
(0042)
(0043)
(00441
(0045)
(0046)
(0047)

52
41
54
48
45
52
00

'R'
'A'

(0040)

06

T
'H'

'E'
'R'
CR

6-2

Flowchart:

Pomter

41

Length

Yes

Length
POinter

= Length

+1

= Pointer

+1

(40)

Length

Source Program:

CHKCR:

LD
LD
LD
CP

HL.41 H

JR

Z.DONE
B
HL
CHKCR

INC
INC

JR
DONE:

LD
LD
HALT

B.O
A.ODH
(HL)

A.B

:POINTER = START OF STRING

:STRING LENGTH = ZERO
:GET ASCII CARRIAGE RETURN TO COMPARE
;IS CHARACTER A CARRIAGE RETURN?
;YES. DONE
:NO. ADD 1 TO STRING LENGTH
:TRY NEXT CHARACTER
;SAVE STRING LENGTH

(40H).A

6-3

Object Program:
Memorv Address
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012

21
41
00
06
00
3E
aD
BE
28
04
04
23
18
F9
78
32
40
00
76

Instruction
(Mnemonic)

CHKCR:

DONE:

LD

HL.41 H

LD

B.O

LD

A.ODH

CP
JR

(HL)
Z.DONE

INC
INC
JR

B
HL
CHKCR

LD

A.B

LD

(40H).A

HALT

The camage return (CR) IS just another ASCII character (hex aD) as far as the computer
IS concerned. The fact that the output device treats the carnage return as a control
character rather than as a pnntlng character does not affect the computer.
The Compare Instruction. CPo sets the flags as if a subtraction had been performed. but
leaves the camage return character In the Accumulator for later comparisons. The Zero
(Z) flag IS affected as follows:
Z= 1

if the character in the stnng is a carnage return

Z= a

if It is not a carnage retu rn

The instruction INC B adds 1 to the string length counter in Register B. LD B.O Initializes
this counter to zero before the loop begins. Remember to initialize vanables before
uSing them in a loop.
This loop does not terminate because a counter is decremented to zero. The computer
will simply continue examining characters until it finds a camage return. You may have
to place a maximum count in a loop like this to avoid problems with erroneous strings
that do not contain a carnage return. What would happen if the example program were
used with such a stnng?
Note that. by rearranging the logiC and changing the initial conditions. you can shorten
the program and decrease ItS execution time. If we adjust the flowchart so that the program increments the counter and pOinter before it looks for the camage return. only one
Jump instruction is necessary instead of two. The new flowchart and program are as
follows:

6-4

Flowchart:

POinter

40
-1

Length

length = Length + 1
POInter = POinter + 1

(40)

Length

Source Program:

CHKCR:

LD
LD
LD
INC
INC
CP

HLAOH
B.OFFH
A.ODH
HL
B

JR

NZ.CHKCR
A.B
(40Hl.A

LD
LD
HALT

(HU

:POINTER = BYTE BEFORE STRING

:LENGTH =-1
:GET ASCII CARRIAGE RETURN TO COMPARE
:ADD 1 TO STRING LENGTH
:IS CHARACTER A CARRIAGE RETURN?
:NO. CHECK NEXT CHARACTER
:YES. SAVE STRING LENGTH

6-5

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010

21
40
00
06
FF
3E
OD
23
04
BE
20
FB
78
32
40
00
76

Instruction
(Mnemonic)

CHKCR:

LD

HLAOH

LD

B.OFFH

LO

AOOH

INC
INC
CP
JR

HL
8
(HL)
NZ.CHKCR

LO
LD

A,B
(40H),A

HALT

The task of looking for a particular value In a list table, or stnng IS a common one. The
Z80 microprocessor has. in fact special instructions that simplify this task.
These special Instructions are called Block Search Instructions;
they operate as follows:

I""':""'~~-""""
BLOCK

SEARCH
CPI compares the contents of the memory location addressed by INSTRUCTIONS ..
HL with the contents of the Accumulator (just like CP (HUl. It then
Increments HL and decrements the byte counter (register pair BC). The Panty/Overflow
bit IS reset if the byte counter is decremented to zero and set otherwise. CPO IS the
same instruction except that It decrements HL Instead of incrementing It.
CPIR and CPDR are the repeated forms of the Block Search instructions. These instructions repeat the basic Search instruction until either BC IS decremented to zero or a true
comparison occurs (i.e.. A = (HUl. Remember that decrementing BC to zero resets the
Panty/Overflow bit while finding a match sets the Zero bit.
Note that BC contains a 16-bit counter. Thus. the Block Search Instructions can handle
stnngs of any length.
A version of the prevIous program uSing CPI IS shown below.
Source Program:

CHKCR:

LD
LD
LD
CPI
JR
LD
SUB
LO
HALT

HLAl H
BC.O
A,ODH
NZ.CHKCR
A.OFFH

:POINTER = START OF STRING

:BYTE COUNTER = ZERO
:GET ASCII CARRIAGE RETURN TO COMPARE
:IS CHARACTER A CARRIAGE RETURN?
:NO, CHECK NEXT CHARACTER
:YES. CALCULATE STRING LENGTH

C
(40H),A

:SAVE STRING LENGTH

6-6

Object Program:
Memory Address

(f~ex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012

21
41
00
01
00
00
3E
00
ED
Al
20
FC
3E
FF
91
32
40
00
76

Instruction
(Mnemonic)

CHKCR:

LO

HL.41 H

LO

BC.O

LO

A.OOH

CPI
JR

NZ.CHKCR

LO

A.OFFH

SUB
LO

C
(40Hl.A

HALT

A little manipulation is necessary to calculate the string length. since CPI decrements
the byte counter (BC) Instead of incrementing It as we did with INC B In the earlier program. Also. the byte counter IS decremented one extra time when the camage return IS
found. How cou Id you adjust the initial conditions to handle this problem?
In fact. we can Imorove the program even further by using CPIR to remove the need for
the relative lump JR. CPIR does everything that CPI does. but it also automatically
repeats the comparison procedure unless A = (HL) or BC has been decremented to zero.
The program uSing CPIR is shown below.

Source Program:
LO
LO
LO
CPIR
LO
SUB
LD
HALT

HL,41 H
BC.O
A.OOH
A.OFFH

:POINTER = START OF STRING

:BYTE COUNTER = ZERO
:GET ASCII CARRIAGE RETURN TO COMPARE
:SEARCH FOR CARRIAGE RETURN
:CALCULATE STRING LENGTH FROM COUNTER

C
(40H).A

:SAVE STRING LENGTH

6-7

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010

21
41
00
01
00
00
3E
00

LO

HL.41 H

LO

BC,O

LO

A,OOH

ED

CPIR

Bl
3E
FF
91
32
40
00
76

Instruction
(Mnemonic)

LO

A,OFFH

SUB
LO

C
(40H),A

HALT

The multiple operation Instructions like CPI and CPIR have the same effect as the sequences they replace. The savings in execution time and memory come about because
the processor needs fewer instructions for each pass through the loop. Thus, the real
savings IS in loop execulon.
All these programs assume that the string is less than 256 bytes long. How would you
change them to handle longer strings?

Find First Non-Blank Character

Purpose: Search a string of ASCII characters (seven bits With most significant bit zero)
for a non-blank character. The string starts in memory location 0042. Place
the address of the first non-blank character into memory locations 0040 and
0041 (most significant bits In 0041), A blank character is hex 20 in ASCII.
Sample Problems:

a.
Resu It:

b.

Result:

(0042)

37

(0040)
(0041)

42, since memory location 0042 contains a non-blank

character.
00

(0042)
(0043)
(0044)
(0045)
(0046)

20
20
20
46
20

(0040)

45, since the three previous memory locations all

contain blanks.
00

(0041)

'7'

SP
SP
SP
F
SP

6-8

Flowchart:

POinter

42

No

POinter =

POinter + 1

(40 and 41)

= POinter

Source Program:

CHBlK:

DONE,

lD
lD
CP
JR
INC
JR
lD

Hl,42H
A,20H
(HU
NZ,DONE
Hl
CHBlK
(40Hl.Hl

:POINTER = START OF STRING

:GET ASCII SPACE FOR COMPARISON
:IS CHARACTER AN ASCII SPACE?
:NO. THROUGH
:YES. EXAMINE NEXT CHARACTER
:NO. SAVE ADDRESS OF FIRST NON-BLANK
CHARACTER

HALT
Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE

21
42
00
3E
20
BE
20
03
23

Instruction
(Mnemonic)

CHBlK,

18

FA
22
40
00
76

DONE,

lD

HL.42H

lD

A,20H

CP
JR

(HU
NZ.DONE

INC
JR

Hl
CHBlK

LD

(40Hl.Hl

HALT

6-9

Looking for spaces in strings IS a common task. Spaces often are eliminated from
strings when they are used simply to increase readability or to fit partlcu lar formats. It is
obviously wasteful to store and transmit beginning. ending or extra spaces. particularly
if you are paying for the communications capability and memory required. Data and
program entry. however. are much simpler if extra spaces are tolerated. Microcomputers are often used In situations like this to convert data between forms that are easy
for humans to use and forms that are efficiently handled on computers and communications lines.
The Instruction LD (addrl.HL is convenient for stonng addresses in the Z80 format (least
significant byte first). LD (40Hl.HL stores the contents of Register L in memory location
0040 and the contents of Register H In memory location 0041.
Again. if we alter the initial conditions so that the loop control section precedes the processing section. we can reduce the number of bytes In the program and decrease the
loop's execution time. The rearranged flowchart is:

Pointer

POinter

f40

and

':=

41

Pointer + 1

41)

0:::

POinter

Source Program:

CH8LK:

LD
LD
INC
CP

JR
LD

HL,41 H
A,20H
HL
(HL)
Z.CHBLK
(40Hl.HL

:POINT TO BYTE BEFORE STRING

:GET ASCII SPACE FOR COMPARISON
;IS CHARACTER .AN ASCII SPACE?
:YES. KEEP EXAMINING CHARACTERS
:NO. SAVE ADDRESS OF FIRST NON-BLANK
CHARACTER

HALT

6-10

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC

21
41
00
3E
20
23
BE
28
FC
22
40
00
76

Instruction
(Mnemonic)

CHBlK:

lD

Hl.41H

lD

A,20H

INC
CP
JR

Hl
(HU
Z.CHBlK

lD

(40H).Hl

HALT

As In the previous example. we could replace the sequence INC Hl, CP (HU with the
single instruction CPI. However. since we do not need the byte counter In this program.
CPI takes Just as much memory (two bytes) and more time (16 clock cvcles Instead of
13) than the instructions It replaces. We could not use CPIR here since we want the program to terminate when the characters are not the same.

Replace Leading Zeros with Blanks

Purpose: Edit a string of ASCII deCimal characters by replacing all leading zeros with
blanks. The string starts in memory location 0041: assume that It consists
entirely of ASCII-coded decimal digits. The length of the string IS in memory
location 0040.
Sample Problems:

(0040)
(0041)

a.

=
=

02
36

'6'

The program leaves the string unchanged, since the leading digit is not zero.
b.

Result:

(0040)
(0041)
(0042)
(0043)

08
30
30
38

'0'
'0'
'8'

(0041)
(0042)

20
20

SP
SP

6-11

Flowchart:

Count

(40)

POinter

41

(Pointer) = ASCII SP

= 20 (Hexl

Pointer

Count

.==

POinter + 1
Count - 1

Source Program:

CHKZ:

DONE:

LD
LD
LD
INC
CP
JR
LD
DJNZ
HALT

HLAOH
B,(HL)
A:O'
HL
(HL)
NZ,DONE
(HL),20H
CHKZ

:COUNT = STRING LENGTH

:GET ASCII ZERO FOR COMPARISON
;15 LEADING DIGIT ZERO?
;NO, THROUGH
;REPLACE LEADING ZERO WITH BLANK
;EXAMINE NEXT DIGIT IF ANY

Single quotation marks around characters indicate ASCII.

6-12

Object Program:

Memory Address
(Hexl

Memory Contents
(Hexl

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
000A
OOOB
OOOC
DODD
OOOE

21
40
00
46
3E
30
23
BE
20
04
36
20
10
F8
76

Instruction
(Mnemonic)

CHKZ:

DONE:

LD

HL.40H

LD
LD

B.(HU
A.-D'

INC
CP
JR

HL
(HU
NZ.DONE

LD

(HU.20H

DJNZ

CHKZ

HALT

You will frequently want to edit decimal strings before they are printed or displayed to
improve their appearance. Common editing tasks include eliminating leading zeros.
JustifYing numbers. adding signs or other identifYing markers. and rounding. Clearly.
printed numbers like 0006 or $27.34382 can be confUSing and annoying.
Here the loop has two exits - one if the processor finds a non-zero digit and the other if
it has examined the entire stnng.
The instruction LD (HU.20H places 20 (hex) Into the memory location addressed by
Registers Hand L. You could also initialize Register C to 20 hex (i.e.. LD C.20H) and use
LD (HU.C to replace the leading zero With a blank. Note the tradeoffs Involved In this example. LD (HU.C executes faster than LD (HU.20H and would thus decrease the Inner
loop's execution time. The overhead reqUired. however. IS an LD C.20H Instruction in
the initialization section of the routine. If this example were to be used in a cash register
application. which sequence would you choose and why?
All digits In the string are assumed to be ASCII: that IS. the digits are hex 30 through 39
rather than the ordinary decimal 0 to 9. The conversion from decimal to ASCII is Simply
a matter of adding hex 30 to the decimal digit.
You may have to be careful. when blanking leading zeros. to leave one zero In the event
that all the digits are zero. How would you do this?
Note that each ASCII digit requires eight bits. as compared to four for a BCD digit.
Therefore. ASCII IS an expensive format in which to store or transmit numerical data.

Add Even Parity to ASCII Characters

Purpose: Add even parity to a stnng of 7-blt ASCII characters. The length of the string
IS in memory location 0040 and the stnng itself begins In memory location
0041. Place even parity In the most significant bit of each character by setting the most significant bit to 1 if that makes the total number of 1 bits in
the word an even number.

6-13

Sample Problem;

Result:

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)
(0046)

06
3,1
32
33
34
35
36

(0041)
(0042)
(0043)
(0044)
(0045)
(0046)

81
82
33
84
35
36

Flowchart;

Pointer

41

Count

(40)

(Pointer! = (Pomter!
OR10000000B
(set parity bit!

POinter

= POinter

Count

= Count -

6-14

+1
1

Source Program:

SETPR:

CHCNT

LD
LD
LD
INC
LD
OR
JP
LD
DJNZ
HALT

HL.40H
B,(HU
:GET STRING LENGTH
C,10000000B :GET PARITY BIT OF 1
HL
A,(HU
:GET A CHARACTER
:SET PARITY BIT TO 1 AND TEST PARITY
C
PO,CHCNT
:IS PARITY NOW EVEN?
:YES. SAVE CHARACTER WITH EVEN PARITY
(HU.A
SETPR

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF

21
40
00
46
OE
80
23
7E
B1
E2
OD
00

Instruction
(Mnemonic)

SETPR:

77
10
F7
76

CHCNT

LD

HL.40H

LD
LD

B.(HU
C,10000000H

INC
LD
OR
JP

HL
A.(HU

LD
DJNZ

(HU.A
SETPR

"-

PO,CHCNT

HALT

Parity IS often added to ASCII characters before they are transmitted on nOIsy communication lines, to provide a simple error-checking facility. Parity detects all single-bit
errors but does not allow error correction (i.e.. you know that an error has occurred
when the received parity IS wrong, but you cannot tell which bit was changed).
LD C, 1OOOOOOOB saves a panty bit of 1 in Register C. (Note the use of the binary mask:
the purpose of the mask IS clearer when It is specified in thiS manner rather than as 80H
or 128 deCimal.)
The instruction OR C sets the panty (most significant) bit to 1 while retaining all the
other bits as they were. as well as selling the Z80 Panty flag.
The follOWing procedure is used to determine if the panty of the byte in memory is odd
or even. We OR a panty bit Into the byte loaded from memory and then test to see if the
panty is odd. If the panty IS odd. then the byte In memory has even panty. and we jump
down to decrement the count of remaining bytes. If the panty is even. then we know
that the byte in memory has odd panty, and therfore we store the byte in the Accumulator Into that memory location.
The conditional jumps JP PO (Jump on Parity Odd) and JP PE (Jump on Panty Even) are
seldom used except In panty generation and checking. Note that there are no relative
jumps conditional on the value of the Panty bit. just as there are none conditional on
the value of the Sign bit.

6-15

Do not confuse the Panty bit 1nluded In each character and the Z80's Parity flag,
which IS set to 1 if the last anthmetlc or 800lean result had even parity.
An alternative approach uses the Z80 SET instruction. This version takes a little longer
but does not require a temporary register for the parity bit.
Source Program:

SETPR:

CHCNT

LD
LD
INC
LD
OR
JP
SET
DJNZ
HALT

HL,40H
B,(HL)
HL
A,(HL)

:GET STRING LENGTH

:GET A CHARACTER
;OOES CHARACTER HAVE EVEN PARITY?

A
PE,CHCNT
7,(HL)
SETPR

:NO, SET PARITY BIT TO 1

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE

21
40
00
46
23
7E
B7
EA
OC
00
CB
FE
10
F6
76

Instruction
(Mnemonic)

SETPR:

CHCNT

LD

HL,40H

LD
INC
LD
OR
JP

8,(HL)
HL
A,(HL)
A
PE,CHCNT

SET

7,(HL)

OJNZ

SETPR

HALT

Pattern Match
Purpose: Compare two stnngs of ASCII characters to see if they are the same. The
length of the strings IS In memory location 0041, one string starts In memory
location 0042 and the other in memory location 0052. If the two stnngs
match, clear memory location 0040: otherwise, set memory location 0040 to
FF hex (all ones).
Sample Problems:
a.

Result:

(0041)

03

(0042)
(0043)
(0044)

43
41
54

(0052)
(0053)
(0054)

43
41
54

'C'
'A'
T

(0040)

00,

since the two strings are the same.

'C'
'A'

6-16

b.

Result:

(0041)

03

(0042)
(0043)
(0044)

52
41
54

(0052)
(0053)
(0054)

43
41
54

(0040)

FF. since the first characters

stnngs differ.

'R'

'A'
T

'C'
'A'
T
In

the

Note: The matching process ends as soon as the CPU finds a difference the strings need not be examined.
Flowchart:

POinter 1 = 42
POinter 2 = 52
Count = (411
Mark = FF (he.1

Mark

(401

Mark

6-17

the rest of

Source Program:

CHCAR:

DONE:

LD
LD
INC
LD
LD
LD
CP
JR
INC
INC
DJNZ
LD
LD
LD
HALT

HL,41H
B.(HU
HL
DE.52H
C.OFFH
A.(DE)
(HL)
NZ.DONE
DE
HL
CHCAR
C.O
A.C
(40Hl.A

:COUNT = LENGTH OF STRINGS

;POINTER 1 = START OF STRING 1
;POINTER 2 = START OF STRING 2
:MARK = FF (HEX)
:GET CHARACTER FROM STRING 2
:IS THERE A MATCH?
NO. DONE

:CHECK NEXT PAIR IF ANY LEFT

:MARK = 0 IF ALL CHARACTERS MATCH
;SAVE MARK

Object Program:
Memorv Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
DODD
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018

21
41
00
46
23
11
52
00
DE
FF
lA
BE
20
06
13
23
10
F8
DE
00
79
32
40
00
76

Instruction
(Mnemonic)

CHCAR:

DONE:

LD

HL.41 H

LD
INC
LD

HL

LD

C.OFFH

LD
CP
JR

A.(DE)
(HL)
NZ.DONE

INC
INC
DJNZ

DE
HL
CHCAR

LD

C.O

LD
LD

A.C
(40Hl.A

B.(HU
DE.52H

HALT

Matching strings of ASCII characters is an essential part of looking for commands.

recognizing names. identifYing vanables or operation codes In assemblers and compilers. finding files. and many other tasks.
The program uses two pointers. one In Register Pair HL and the other In Register Pair
DE. The onlv instructions that use the address In DE are LD A.(DE) (Load Accumulator
From Memory Location Addressed by DE) and LD (DEl.A (Store Accumulator in Memory
Location Addressed by DE). ArithmetiC and logical operations With memory and
transfers to or from other registers (e.g.. ADD A(HL); AND (HL); LD B. (HL); LD (HL).E)
can only be performed using the address in Register Pair HL. or using an Index register.

6-18

The order of operations is very important because of the small number of instructions
that use the address In Register Pair DE. You must move a character from the string
pOinted to by DE to the Accumulator and compare It to a character in the string pointed
to by HL. This order of operations IS necessary because the Z80 has no instruction
which allows a comparison to a character in a string pOinted to by DE.
For example. if you replaced LD A.{DE) With LD A,(HL). what would the next instruction
be? This asymmetry IS peculiar to the Z80 and can cause programming nightmares.
Note that each Iteration updates both pointers.
ThiS program could take advantage of the fact that a register IS known to contain zero
after a particular conditional Jump is executed. When the DJNZ CHCAR instruction is
executed. if the branch IS not performed. then we know that Register B contains zero.
Therefore. we can move Register B to Register C. our flag register. to indicate that a
match has been found.
We could also use the Z80's SET and RESET Instructions to handle the flag if we
needed to conserve bits for other purposes.

PROBLEMS
1) Length of a Teletypewriter Message
Purpose: Determine the length of an ASCII message. All characters are 7-bit f\SCII
With MSB = O. The stnng of characters In which the message IS embedded
starts In memory location 0041. The message Itself starts With an ASCII STX
character (hex 02) and ends with ETX (hex 03). Place the length of the
message (the number of characters between the STX and the ETX but including neither) Into memory location 0040.
Sample Problem:

Result:

2)

(0041)
(0042)
(0043)
(0044)
(0045)

40
02
47
4F
03

(0040)

02. since there are two characters between

the STX In location 0042 and ETX in
location 0045.

STX
'G'
'0'
ETX

Find Last Non-Blank Character

Purpose: Search a string of ASCII characters for the last non-blank character. The
string starts in memory location 0042 and ends With a carriage return
character (hex 00). Place the address of the last non-blank character into
memory locations 0040 and 0041 (most significant bits in 00411.
Sample Problems:

a.
Result:

(0042)
(0043)

37
00

(0040)

42. since the last (and only) non-blank character

IS In memory location 0042.
00

(0041)

'7'
CR

6-19

b.

Result:

3)

(0042)
(0043)
(0044)
(0045)
(0046)
(0047)
(0048)
(0049)

ob

(0040)
(0041)

46
00

41
20
48
41
54
20
20

'A'
SP
'H'
'A'
'T'
SP
SP
CR

Truncate Decimal String to Integer Form

Purpose: Edit a string of ASCII decimal characters by replacing all digits to the right of

the decimal POint with ASCII blanks (hex 20). The string starts In memory
location 0041 and is assumed to consist entirely of ASCII-coded decimal
digits and a possible decimal point (hex 2E). The length of the string is in
memory location 0040. If no decimal point appears in the string, assume that
the decimal point is implicitly at the far right.
Sample Problems:

a.

Result:

b.

(0040)

04

(0041)
(0042)
(0043)
(0044)

37 '7'
2E"
38 '8'
31 '1'

(0041)
(0042)
(0043)
(0044)

37'7'
2E"
20 SP
20 SP

(0040)

03

(0041)
(0042)
(0043)

26 '6'
37 '7'
31 '1'
Unchanged, as number is assumed to be 671.

Result:

4)

Check Even Parity in ASCII Characters

Purpose: Check even parity

In a stnng of ASCII characters. The length of the string IS

memory location 0041, and the stnng Itself begins in memory location
0042. If the parity of all the characters In the string is correct. clear memory
location 0040: otherwise, place FF hex (all ones) into memory location 0040.
In

6-20

Sample Problems:
a.

Result:
b.

Result:

5)

(0041)

03

(0042)
(0043)
(0044)

B1
B2

33

(0040)

00,

(0041)

03

(0042)
(0043)
(0044)

B6
33

(0040)

since all the characters have even parity.

B1

FF since the character In memory

location 0042 does not have even parity.

String Comparison

Purpose: Compare two strings of ASCII characters to see which is larger (i.e.. which
follows the other in 'alphabetical' ordering). The length of the strings is In
memory location 0041, one string starts in memory location 0042 and the
other in memory location 0052. If the string starting In memory location
0042 is greater than or equal to the other string, clear memory location
0040: otherwise, set memory location 0040 to FF hex (all ones).
Sample Problems:

a.

Result:
b.

Result:

c.

Result:

(0041)

03

(0042)
(0043)
(0044)

43
41
54

'C'

(0052)
(0053)
(0054)

42
41
54

'B'
'A'
T

(0040)

00,

since CAT

(0041)

03

(0042)
(0043)
(0044)

43
41
54

(0052)
(0053)
(0054)

43
41
54

(0040)

00,

since the two strings are equal.

(0041)

03

(0042)
(0043)
(0044)

43
41
54

'C'
'A'

(0052)
(0053)
(0054)

43
55
54

'C'
'U'
T

(0040)

FF.

since CUT

'A'
T

IS

'larger thitn BAT

'C'

'A'

T
'C'

'A'

6-21

IS

'larger' than CAT

Chapter 7
CODE CONVERSION
Code conversion is a continual problem in most microcomputer applications. Peripherals provide data In ASCII. BCD. or various special codes. The system must convert the
data into some standard form for processing. Output devices may require data In ASCII.
BCD. seven-segment. or other codes. Therefore, the system must convert the results to
a sUitable form after the processing IS completed.
There are several ways to approach code conversion:
1)

Some conversions can easily be handled by algorithms Invo[ving arithmetic or [oglcal functions. The program may, however, have to handle some special cases separately.

2)

More complex conversions can be handled With lookup tables. The lookup table
method requires [ittle programming and IS easy to apply. However. the table may
occupy a large amount of memory if the range of input values IS large.

3)

Hardware IS readily available for some conversion tasks. Typlca[ examples are
decoders for BCD to seven-segment conversion and Universal Asynchronous
Receiver/Transmitters (UARTs) for conversion between para[lel (ASCII) and serial
(teletypeWriter) formats.

In most applicatIOns, the program should do as much as possible of the code conversion
work. ThiS results In a savings in parts and board space as we[1 as In increased
reliability. Furthermore, most code conversions are easy to program and require [ittle
execution lime.

EXAMPLES
Hex to ASCII
Purpose:

Convert the contents of memory location 0040 to an ASCII character.

Memory location 0040 contains a Single hexadecimal digit (the four most
significant bits are zero). Store the ASCII character In memory location
0041.

Sample Problems:
a.
Result:
b.
Result:

(0040)

OC

(0041)

43

(0040)

06

(0041)

36

'C'

'6'

7-1

Flowchart:

Data =(401

Ves

Data

= Data
+ ASCII A
-ASCII 9 - 1

Result =
Data + ASCII Zero

(411 = Result

Source Program:

ASCZ:

LD
CP
JR
ADD
ADD
LD
HALT

A,(40H)
10
C,ASCZ
A.'A'-'9'-1
A.'O'
(4IHl,A

:GET DATA
:IS DATA 10 OR MORE?
:YES,ADD OFFSET FOR LETTERS
:ADD OFFSET FOR ASCII
:STORE ASCII RESULT

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
0008
OOOC
DODD
OOOE

3A
40
00
FE
OA
38
02
C6
07
C6
30
32
41
00
76

Instruction
(Mnemonic)

ASCZ:

LD

A,(40H)

CP

10

JR

C,ASCZ

ADD

A.'A'-'9'-1

ADD

A.'O'

LD

(41Hl,A

HALT

7-2

In this program. the basIc Idea IS to add ASCII 0 to all the hexadecimal digits. This addition converts the decimal digits correctly; however. there IS a break between ASCII 9
(39 hex) and ASCII A (41 hex) which must be considered. This break must be added to
the nondecimal digits A, 8. C. D. E, and F This is accomplished by the ADD A instruction which adds the offset 'A'-'9'-1 to the contents of the Accumulator. Can you explain
why the offset IS .A' -'9'-17
Note that the addition terms are placed in the assembly language program in ASCII
form (apostrophes surround an ASCII character or string of characters). The offset for
the letters is left as an arithmetic expression. The effort IS to make the purpose of the
terms as clear as possible in the assembly language listing. The extra assembly time IS a
very small price to pay for a large Increase in clarity.
This routine could be used In a variety of programs; for example, monitor programs
must convert hexadecimal digits to ASCII In order to display the contents of memory
locations in hexadecimal on an ASCII printer or video display.
Another (quicker! conversion method that requires no conditional lumps at all is the
following program, described by Allison In Computer magazine. 1
LD
ADD
DAA
ADC
DAA
LD
HALT

A,(40HI
A,90H

;GET HEX LJIGIT

;DEVELOP EXTRA 6 AND CARRY

AAOH

;ADD IN CARRY. ASCII OFFSET

(4IHI,A

;STORE ASCII DIGIT

Try this program on some digits. Can you explain why It works?

Decimal to Seven-Segment
Purpose:

Convert the contents of memory location 0040 to a seven-segment code In

memory location 0042. If memory location 0040 does not contain a Single
deCimal digit. clear memory location 0042.

Seven-segment table; The following table can be used to convert decimal numbers to
seven-segment code. The seven-segment code is organized With the most significant
bit always zero followed by the code (1 = on, 0 = offl for segments g, f, e, d, c, b. and a
(see Figure 7-1).

Digit

Code

0
1
2
3
4

3F
06

5
6
7

8
9

58
4F
66
6D
7D
07
7F
6F

Figure 7-1. Seven-segment Arrangement

7-3

Note that the table uses 70 for 6 rather than the alternative 7C (top bar off) to avoid
confusion with lower case b. and 6F for 9 rather than 67 (bottom bar off), for no particular reason.
Sample Problems:
(0040)

a.
Result:
b.
Result:

03

(0042)

4F

(0040)

28

(0042)

00

Flowchart:

Data =(40)

Ves

Result

Result =0

(SSEG + Datal

(421

= Result

Note that the addition of base address SSEG and Index (DATA) produces the address
that contains the answer.

7-4

Source Program:

LD
LD
CP

JR

DONE.

SSEG:

LD
LD
LD
ADD
LD
LD
LD
HALT
ORG
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB

B.O
A.(40H)
10
NC.DONE
L.A
H.O
DE.SSEG
HL.DE
B.(HL)
A,B
(42Hl.A

;GET ERROR CODE TO BLANK DISPLAY

:GET DATA
;IS DATA A DECIMAL DIGIT?
:NO. KEEP ERROR CODE
;YES. MAKE DATA INTO A 16-BIT INDEX

20H
3FH
06H
5BH
4FH
66H
6DH
7DH
07H
7FH
6FH

:SEVEN-SEGMENT CODE TABLE

;GET BASE ADDRESS OF 7-SEGMENT TABLE

;FIND ELEMENT BY INDEXING
:GET 7-SEGMENT CODE FROM TABLE
:SAVE 7-SEGMENT CODE OR ERROR CODE

7-5

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015

06
00
3A
40
00
FE
OA
30
08
6F
26
00
11
20
00
19
46
78
32
42
00
76

0020
0021
0022
0023
0024
0025
0026
0027
0028
0029

3F
06
5B
4F
66
60
70
07
7F
6F

Instruction
(Mnemonic!

DONE:

LD

B,O

LD

A,(40H)

CP

10

JR

NC,DONE

LD
LD

L,A
H,O

LD

DE,SSEG

ADD
LD
LD
LD

HL,DE
B,(HU
A.B
(42Hl.A

HALT
SSEG:

DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB

3FH
06H
5BH
4FH
66H
6DH
7DH
07H
7FH
6FH

The program calculates the memory address of the desired code by adding the Index
(i.e.. the digit to be displayed) to the base address of the seven-segment code table.
This procedure IS known as a table lookup.
The assembly language pseudo-opera lion DEFB (Define Byte) places constant data into
program memory. Such data may Include tables, headings, error messages, priming
messages. format characters, thresholds, etc. The label attached to a DEFB pseudooperation IS assigned the value of the address Into which the byte of data is placed.
Tables are often used to perform code conversions that are more complex than the previous example. Such tables typically contain all the results organized according to the
input data, e.g.. the first entry IS the code corresponding to the number zero.
Seven-segment displays provide recognizable forms of the decimal digits and a few letters and other characters. Calculator-type seven-segment displays are inexpensive,
easy to combine, and use little power. However. the seven-segment coded digits are
somewhat difficult to read.
The assembler simply places the data for the table into memory. Note that one DEFB
pseudo-operation fills one byte of memory. We have left some memory space between
the program and the table to allow for later additions or corrections.

7-6

An alternative approach would be to use one of the Z80's Index

registers, say IX. The programmer must be aware of the following
features of the Z80's Index registers:

USE OF zao
INDEX
REGISTERS

1)

The fixed offset In program memory IS only eight bits long and
so cannot hold a complete memory address. It must be used either as a short displacement or to hold the eight least Significant bits of a memory address.

2)

The Index registers are 16 bits long. Either IX or IY can be loaded from memory Just
like a register pair - from two consecutive memory addresses with the least significant eight bits at the lower address.

3)

All operations involVing the Index registers take extra time and memorv because
one word of the operation code Simply declares that an Index register is to be used.

The following program uses Register IX to perform the table lookup:

Source Program:
LD
LD
CP
JR
LD
LD
LD
LD
DONE.
LD
LD
HALT

B,O
A,(40H)
10
NC,DONE
HL,41 H
(HLI,O
IX,(40H)
B,(IX+SSEG)
A,B
(42Hl.A

:GET ERROR CODE TO BLANK DISPLAY

:GET DATA
:IS DATA A DECIMAL DIGIT?
:NO, KEEP ERROR CODE
:SAVE TABLE PAGE NUMBER IN MEMORY
:GET TABLE OFFSET
:GET 7-SEGMENT CODE FROM TABLE
:SAVE 7-SEGMENT CODE OR ERROR CODE

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019

06
00
3A
40
00
FE
OA
30
OC
21
41
00
36
00
DO
2A
40
00
DO
46
20
78
32
42
00
76

Instruction
(Mnemonic)

DONE.

LD

B.O

LD

A.(40H)

CP

10

JR

NC.DONE

LD

HL,41 H

LD

(HLI,O

LD

IX,(40H)

LD

B,(IX+SSEG)

LD
LD

A.B
(42Hl.A

HALT

7-7

The indexed load mstruction LD B.OX + SSEG) adds the index (j.e.. the digit to be displayed) to the base of the seven-segment table to get the address of the desired code.
Note that the 16-bit index register contains the data as its eight least significant bits
and the most significant bits of the starting address of the table as Its eight most significant bits. This odd arrangement is necessary because the offset Included with the mdexed instruction is only eight bits long and can therefore hold only the eight least significant bits of the starting address of the table.
A more general program would allow the table to be placed anywhere in memory. If the
table startmg address is SSEGM (eight MSBs) and SSEGL (eight LSBs). the instruction
LD (HU.O must be replaced by LD (HU.SSEGM. Why is thiS change necessary?
Note that all operations involvmg Index Register IX have a 2-word operation code in
which the first word is DD.
Clearly thiS IS not a very efficient use of the mdex registers. These
MOVING DATA
WITHIN
registers really become useful when you must access several data
in a block. The block might contain the characteristics of a
A BLOCK
message. the parameters of an equation. the current state of a process or machine. or the data for a video display. You could, for example. take the COntents of the twelfth location m the block and move them to the twentieth location With
either of the followmg programs. assuming that the starting address of the block IS
stored in memory locations PTR and PTR+l.
1)

Using DE and HL.

LD
LD
ADD
LD
LD
ADD
LD

2)

DE.(PTR)
HL.12
HL.DE
A,(HU
HL.20
HL,DE
(HU.A

;GET STARTING ADDRESS

:CALCULATE SOURCE ADDRESS

IX.(PTR)
A.OX+12)
OX+201.A

:GET STARTING ADDRESS

;GET DATA FROM SOURCE
;MOVE DATA TO DESTINATION

:GET DATA FROM SOURCE

:CALCULATE DESTINATION ADDRESS
;MOVE DATA TO DESTINATION

Usmg IX.
LD
LD
LD

The program using the mdex registers IS far shorter and clearer. Its only limitation is
that the offsets must be small enough to fit into an 8-blt byte.

ASCII to Decimal
Purpose: Convert the contents of memory location 0040 from an ASCII character to a
decimal digit and store the result in memory location 0041. If the contents of
memory location 0040 are not the ASCII representation of a decimal digit.
set the contents of memory location 0041 to FF (hex).
Sample Problems:

a.

(0040)

37

Result:

(0041)

07

(0040)

55

Result:

(0041)

FF

b.

'7'

7-8

Flowchart:

Data =(401

Yes

Yes

Result

=
Result = FF (Hexl

Data - ASCII 0

(411 = Result

Source Program:

LD
LD
SUB

JR

DONE:

B.OFFH
A, (40H)

'0'
C.DONE

CP

'9'+1

JR

NC.DONE
B.A
A.B
(41H).A

LD
LD
LD
HALT

:GET ERROR MARKER

:GET DATA
:IS DATA BELOW ASCII ZERO?
:YES. NOT A DIGIT
:IS DATA ABOVE ASCII NINE
:YES. NOT A DIGIT
;SAVE DIGIT IF VALID
:SAVE DIGIT OR ERROR MARKER

7-9

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012

06
FF
3A
40
00
D6
30
38
05
FE
3A
30
01
47
78
32
41
00
76

Instruction
(Mnemonic)

DONE:

LD

B,OFFH

LD

A,(40H)

SUB

'0'

JR

C,DONE

CP

'9'+1

JR

NC,DONE

LD
LD
LD

B,A
A.B
(41H),A

HALT

This program handles ASCII-coded characters lust like ordinary numbers. Note that the
decimal digits and the letters form groups of consecutive codes. Strings of letters (like
names) can be alphabetized by placing their ASCII representations In increasing
numerical order (ASCII B = ASCII A + 1 for example),
Subtracting ASCII zero (30 hex) from any ASCII decimal digit gives the BCD representation of that digit
ASCII to decimal conversion IS necessary when decimal numbers are being entered
from an ASCII device like a teletypewnter or video terminal.
The basIC idea of the program IS to determine if the character is between ASCII 0 and
ASCII 9, inclusive. If the character IS, It'S an ASCII decimal digit. since the digits form a
sequence. It may then be converted to decimal simply by subtracting hex 30 (ASCII 0),
e.g.. ASCII 7 - ASCII 0 = 37-30 = 7
Note that one comparison IS done with an actual subtraction (SUB '0') since the subraction is necessary to convert ASCII to decimaL The other companson is done with an Implied subtraction (CP '9'+1) since the final result IS now in the Accumulator if the onginal number was valid.

BCD to Binary
Purpose: Convert two BCD digits in memory locations 0040 and 0041 to a binary
number in memory location 0042. The most significant BCD digit is in
memory location 0040.
Sample Problems:

a.

(0040)
(0041)
Resu It

b.

Result:

02
09

= 29

(0042)

1D (hex)

(0040)
(0041)

07
01

(0042)

47 (hex) = 71 (deCimal)

(declmaJ)

7-10

Note: No flowchart is Included since the program multiplies the most significant digit
by 10 simply by using the formula lOx = 8x + 2x. Multiplying by 2 requires one
arithmetic left shift and multiplying by 8 requires three such shifts.
Source Program:

LD
LD
ADD
LD
ADD
ADD
ADD
INC
ADD
INC
LD
HALT

HLAOH
A(HU
A.A
B.A
AA
A.A
A.B
HL
A(HU
HL
(HU.A

:GET MOST SIGNIFICANT DIGIT (MSD)

;MSD TIMES TWO
:SAVE MSD TIMES TWO
;MSD TIMES FOUR
;MSD TIMES EIGHT
:MSD TIMES TEN
:POINT TO LEAST SIGNIFICANT DIGIT
;ADD TO FORM BINARY EQUIVALENT
:STORE BINARY EQUIVALENT

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD

21
40
00
7E
87
47
87
87
80
23
86
23
77

76

Instruction
(Mnemonic)
LD

Hl,40H

LD
ADD
LD
ADD
ADD
ADD
INC
ADD
INC
LD
HALT

A(HL)
A.A
B.A
A.A
A.A
A.B
HL
A.(HL)
HL
(HU.A

BCD entries are converted to binary in order to save on storage and to simplify calculations. However. the conversion may offset some of the advantages of binary storage
and arithmetic.
This program multiplies the BCD digit in memory locatIOn 0040 by ten uSing repeated
additions 2 Note that ADD A.A multiplies the contents of the Accumulator by 2. This
allows you to multiply the contents of the Accumulator by small decimal numbers In a
few instructions. How would you use this procedure to multiply by 167 by 12? by 7?
BCD numbers reqUire about 20% more storage than do binary numbers. Representing 0
to 999 reqUires 12 bits In BCD form but only 10 bits In binary (since
2 10 = 1024::::: 1000).

Convert Binary Number to ASCII String

Purpose: Convert the 8-bit binary number In memory location 0041 to eight ASCII
characters (either ASCII 0 or ASCII 1) in memory locatIOns 0042 through
0049 (the most significant bit is in 0042).

7-11

Sample Problem:

Result:

(0041)

02

(0042)
(0043)
(0044)
(0045)
(0046)
(0047)
(0048)
(0049)

31
31
30
31
30
30
31
30

= 11010010

'1'
'1'
'0'
'1'
'0'
'0'
'1'
'0'

Flowchart:

(Pointer)

Counter

=
=
=

POinter

POinter H

Pointer
Data

(Pointer! =

41
8

ASCII 0

Shift Data left one bit

No

(Pointerl
Counter

= Counter -1

ASCII 1. i.e..
(Pointed + 1

Source Program:

CONV:

COUNT

LD
LD
LD
LD
INC
LD
RLA
JR
INC
DJNZ
HALT

HL.41 H

A.(HU
B,8
CO'
HL
(HU.C
NC.COUNT
(HU
CONV

:GET DATA
:COUNTER = NUMBER OF BITS IN WORD
:GET ASCII ZERO TO STORE IN STRING
:PUT ASCII ZERO IN STRING
:IS NEXT 81T OF DATA 1?
:YES. MAKE STRING ELEMENT ASCII ONE

7-12

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010

21
41
00
7E
06
08
OE
30
23
71
17
30
01
34
10
F8
76

Instruction
(Mnemonic)

'CONV:

COUNT

LD

HL,41 H

LD
LD

A.(HU
B.8

LD

CO'

INC
LD
RLA
JR

HL
(HU.C

INC
DJNZ

(HL)
CONY

NC.COUNT

HALT

The ASCII digits form a sequence so ASCII 1 = ASCII 0+ 1. Remember that the Z80
registers have special uses. You should place the loop counter into Register B so that
you can use the DJNZ instruction.
Be careful of the difference between INC HL. which adds one to the 16-blt contents of
Register Pair HL. and INC (HL). which adds one to the 8-blt contents of the memory
location addressed by Register Pair HL.
Binary-to-ASCII conversion is necessary when numbers are printed in binary form on an
ASCII device.
The conversion to ASCII simply Involves adding ASCII 0 (hex 30).

PROBLEMS
1) ASCII to Hex
Purpose: Convert the contents of memory location 0040 to a hexadeCimal digit and
store the result in memory location 0041. Assume that memory location
0040 contains the ASCII representation of a hexadeCimal digit (7 bits with
MSB OJ.
Sample Problems:

a.

(0040)
Resu It:

b.
Result:

2)

43

(0041)

OC

(0040)

36

(0041)

06

'C'

'6'

Seven-Segment to Decimal

Purpose: Convert the contents of memory location OU40 from a seven-segment code
to a deCimal number In memory location 0041. If memory location 0040 does
not contain a valid seven-segment code. set memory location 0041 to FF
(hex). Use the seven-segment table given under the DeCimal to Seven-Segment example and try to match codes.

7-13

Sample Problems:

a.
Resu It:
b
Result:

3)

(0040)

4F

(0041)

03

(0040)

28

(0041)

FF

Decimal to ASCII

Purpose: Convert the contents of memory location 0040 from a decimal digit to an
ASCII character and store the resu It in memory location 0041. If the number
In memory location 0040 is not a deCimal digit. set the contents of memory
location 0041 to an ASCII blank character (20 hex!.
Sample Problems:

(0040)

07

(0041)

37

(0040)

55

(0041)

20

a.
Result:
b.
Result:

4)

'T

SP

Binary to BCD

Purpose: Convert the contents of memory location 0040 to two BCD digits in memory
locations 0041 and 0042 (most Significant digit In 0041). The number In
memory location 0040 IS unsigned and less than 100.
Sample Problems:

a.

(0040)

10

Result:

(0041)
(0042)

02
09

(0040)

47

Resu It:

(0041)
(0042)

07
01

b.

5)

(29 decimal)

(71 deCimal)

ASCII String to Binary Number

Purpose: Convert the eight ASCII characters in memory locations 0042 through 0049
to an B-bit binary number In memory location 0041 (the most Significant bit
IS in 0042). Clear memory location 0040 if all the ASCII characters are either
ASCII 1 or ASCII 0 and set it to FF otherwise.
Sample Problems:

(0042)
(0043)
(0044)
(0045)
(0046)
(0047)
(0048)
(0049)

31
31
30
31
30
30
31
30

(0041)
(0040)

02
00

b. same as 'a' except:

(0045)

37

(0040)

FF

a.

Result:

Result:

'1'

'1'
'0'
'1'
'0'
'0'

'1'
'0'

'7'

7-14

REFERENCES
1.

Allison, D.R.. "A Design Philosophy for Microcomputer Architectures," Computer.

February 1977. pp. 35-41. This IS an excellent article which we recommend highly.

2.

Other BCD-to-blnary conversion methods are discussed in JA Tabb and M.L.

Roginsky, "Microprocessor Algorithms Make BCD-Binary Conversions Super-fast:'
EON. January 5, 1977. pp. 46-50 and In J.B. Peatman, Microcomputer-based
Design, McGraw-HilL New York, 1977. pp. 400-406.

7-15

Chapter 8
ARITHMETIC PROBLEMS
Most arithmetic in microprocessor applications consists of multiple-word binary or
decimal manipulations. A decimal correction (decimal adjust) or some other means for
performing decimal arithmetic is frequently the only arithmetic Instruction provided
besides basic addition and subtraction. You must Implement other arithmetic operations with sequences of instructions.
Multiple-precIsion binary arithmetic requires simple repetitions of the basIc single-word
Instructions. The Carry bit transfers Information between words. Add with Carry and
Subtract with Carry use the Information from the prevIous arithmetic operations. You
must be careful to clear the Carry before operating on the first words (obviously there IS
no carry Into or borrow from the least significant bits).
Decimal arithmetic is a common enough task for microprocessors that most have
special instructions for this purpose. These instructions may either perform decimal
operations directly or correct the results of binary operations to the proper decimal
form. Decimal arithmetic IS essential in such applications as point-of-sale terminals.
calculators. check processors. order entry systems. and banking terminals.
You can implement multiplication and divIsion as series of additions and subtracilons
respectively. much as they are done by hand. Double-word operations are necessary
since a multiplication produces a result twice as long as the operands. while a division
similarly contracts the length of the result. Multiplications and divisions are time-consuming when done In software because of the repeated arithmetiC and shift operations
that are necessary. Of course. multiplying or dividing by a power of 2 is simple because
such operalions can be implemented with an appropriate number of left or right
arithmetiC shifts.

EXAMPLES
Multiple-Precision Addition
Purpose: Add two multiple-word binary numbers. The length of the numbers (in bytes)
is In memory location 0040. the numbers themselves start (least significant
bits first) in memory locations 0041 and 0051. respectively. and the sum
replaces the number starting in memory location 0041.

8-1

Sample Problem:

Result:

that

(0040)

04

(0041)
(0042)
(0043)
(0044)

C3
A7
58
2F

(0051)
(0052)
(0053)
(0054)

88
35
DF
14

(0041)
(0042)
(0043)
(0044)

78
DD
3A

44

IS.

2F58A7C3
14DF3588
443ADD78

Flowchart:

CoUnt

Pointer 1
Pointer 2
Carry

(Pointer 1) =
(Pointer 1) t
(Pointer 2) +

ll-_-.;c;;arrv;.;._ _..1
Pointer 1= Pointer 1
+ 1
POInter 2 = Pointer 2
+ 1

Count

= Cnunt - 1

8-2

(This step also produces a new Carry)

Source Program:

ADDW:

LD
LD
INC
LD
AND
LD
ADC
LD
INC
INC
DJNZ
HALT

HLAOH
B.(HU
HL
DE.51H

A
A.(DE)
A.(HU
(HU.A
DE
HL
ADDW

;COUNT = LENGTH OF STRINGS (IN BYTES)

;POINTER 1 = FIRST WORD OF STRING 1
;POINTER 2 = FIRST WORD OF STRING 2
;CLEAR CARRY TO START
;GET WORD FROM STRING 2
;ADD WORD FROM STRING 1
;STORE RESULT IN STRING 1

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010

21
40
00
46
23
11
51
00
A7
lA
8E
77
13
23
10
F9
76

Instruction
(Mnemonic)

ADDW;

LD

HL.40H

LD
INC
LD

B.(HU
HL
DE.51H

AND
LD
ADC
LD
INC
INC
DJNZ

A
A.(DE)
A.(HL)
(HU.A
DE
HL
ADDW

HALT

The relative address for DJNZ ADDW IS:

09
-10

09
+FO

f.'9
The Instruction AND A is used to clear the Carry bit. Any other logical operation would
have the same effect. The Carry must be cleared. since there is no carry Involved in the
addition of the least significant bytes.
The instruction ADC. Add with Carry. includes the Carry from the previous words In the
addition.ADC IS the only Instruction In the loop that affects the Carry Remember that
neither INC nor DJNZ does.
Both the pOinter In Register Pair DE and the one in HL must be updated during each
Iteration.

8-3

This procedure can add binary numbers of up to 256 bytes In

length. Note that the ten binary bits correspond to three decimal
digits. since 2 10 = 1024
1000. So. you can calculate the number of bits required to give a certain accuracy in decimal digits. For
example. ten decimal digit accuracy requires:

DECIMAL
ACCURACY
IN BINARY

."
10d'"
bitS) = 33 bits
(10 digits)
"x ( 3
Iglts
BLOCK
If we were only transferring the data from one place In memory
TRANSFER
to another and not also processing it. we could use the ZBO's
INSTRUCTIONS
powerfu I block transfer instruction LDIR. This single instruction
moves a byte of data from the address In HL to the addre"ss in
DE. increments the pOinters in HL and DE. and decrements the byte counter in Be. It
repeats the move operation until BC is decremented to zero. LDI is the same instruction
without the repetition factor: LDD and LDDR are non-repeated and repeated moves.
respectively. that decrement the pointers rather than incrementing them.

A program to transfer a fixed number of bytes (LENGTH) from one place in memory
(starting at PTR1) to another place in memory (starting at PTR2) IS the following.

Block Move
Purpose: Move a block of data BC characters long from the address
dress In DE.

In

HL to the ad-

Sample Problem:

Result:

(HL)
(DE)
(BC)

40
50
3

(0040)
(0041)
(0042)

31
32
33

(0050)
(0051)
(0052)

0
0
0

(0050)
(0051)
(0052)

31
32
33

Source Program:

LD
LD
LD

BC.LENGTH
HL.PTR1
DE.PTR2

:COUNT = LENGTH OF TRANSFER (IN BYTES)

:POINTER 1 = START OF DATA SOURCE AREA
:POINTER 2 = START OF DATA DESTINATION
. AREA

LDIR
HALT

8-4

Object Program:

Memory Address
(Hex)
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB

Memory Contents
(Hex)

Instruclion
(Mnemonic)

01

LD

BC.LENGTH

LENGTH
21

LD

HL.PTRl

PTRl
11

LD

DE.PTR2

PTR2
ED
BO
76

LDIR
HALT

Try to Implement the same program without the LDIR instruction. How many bytes of
memory and clock cycles does It require each way?

Decimal Addition
Purpose: Add two multiple-word decimal (BCD) numbers. The length of the numbers
is in memory location 0040. the numbers themselves start (least significant
bits first! In memory locations 0041 and 0051. respectively. and the sum
replaces the number starting in memory location 0041.
Sample Problem:

Result:

that is.

(0040)

04

(0041)
(0042)
(0043)
(0044)

85
19
70
36

(0051)
(0052)
(0053)
(0054)

59
34
66
12

(0041)
(0042)
(0043)
(0044)

44
54
36
49
36701985
+12663459
49365444

8-5

Flowchart:

Count

140)

Pointer 1
POInter 2

41
51

Carry

Pointer 1 =
(PoInter 11 +
(Pointer 2) +

Carry +
Decimal correction

(This step also produces a new Carry'

Pointer 1
POinter 1 + 1

Pointer 2

Pointer 2 + 1

Count

Count - 1

Source Program:

DECAD

LD
LD
INC
LD
AND
LD
ADC
DAA
LD
INC
INC
DJNZ
HALT

HLAOH
B.(HL)
HL
DE.51H

A
A(DE)
A.(HL)
(HL).A
DE
HL
DECAD

:COUNT = LENGTH OF STRINGS (IN BYTESI

:POINTER 1 = FIRST WORD OF STRING 1
:POINTER 2 = FIRST WORD OF STRING 2
:CLEAR CARRY TO START
:GET 2 DECIMAL DIGITS FROM STRING 2
:ADD PAIR OF DIGITS FROM STRING 1
:MAKE ADDITION DECIMAL
:STORE RESULT IN STRING 1

8-6

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
OOOB
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011

21
40
00
46
23
11
51
00
A7
lA
SE
27
77
13
23
10
FS
76

Instruction
(Mnemonic)

DECAD:

LD

HLAOH

LD
INC
LD

B,(HL)
HL
DE.51H

AND
LD
ADC
DAA
LD
INC
INC
DJNZ

A
A,(DE)
A,(HL)

HALT

The Decimal Adjust instruction (DAA) uses the Carry (C) and Half
Carry (H) bits to correct the following situations:
1

(HL),A
DE
HL
DECAD

DECIMAL
ADJUST

The sum of two digits IS between 10 and 15, inclUSive. In this

case, SIX must be added to the sum to give the right result. I.e.
0101

(5)

+ 1000 (S)
1101 (D)
0110
0001 0011 (BCD 13, which IS correctl

2)

The sum of two digits IS 16 or more. In this case the result IS a proper BCD number
but SIX less than It should be, I.e.
1000 (S)
+ 1001 (9)
0001 0001 (BCD 11)
+ 0110
0001 0111 (BCD 17. which IS correctl

Six must be added in both situations. However, case 1 can be recognized by the fact
that the sum IS not a BCD digit. It IS between 10 and 15 (or A and F hexadecimal). Case
2 can be recognized only by the fact that the Carry (most Significant digltl or Half Carry
(least Significant digit) has been set to 1, since the result is a valid BCD number. DAA is
the only instruction that uses the Half Carry. Note that DAA operates only on the Accumulator.

zao

ADD/SUBTRACT
The
microprocessor also has a flag that distinguishes beFLAG
tween Add Instructions (ADD. ADC) and Subtract instructions
(SUB, SBC). This flag, called the Add/Subtract flag or N flag, IS
cleared by all Add instructions and set by all Subtract Instructions. The sole use of this
flag IS to allow the DAA instructIOn to correctly change binary addition Into BCD addition and binary subtraction into BCD subtraction. The SOSO and SOS5 microprocessors
do not have an N flag, and so their DAA Instructions operate properly only after addition.

8-7

DAA can be used only after instructions that place their result into the Accumulator
and that properly affect the Carry. Half-Carry. and Add/Subtract flags. Thus. you cannot
use DAA after INC (since INC does not affect the Carry!. DEC. or any of the double-word
instructions that place their results into the Index registers or Register Pair HL.
ThiS procedure can add deCimal (BCD) numbers of any length.
Here four binary bits are required for each decimal digit. so tendigit accuracy requires:

ACCURACY IN
BINARY AND
BCD
"OJ

10 x 4 = 40 bits
as opposed to 33 bits In the binary case. ThiS IS essentially fiye 8-bit words Instead of
four. The deCimal procedure also takes a little longer per word because of the extra
DAA instruction.

8-Bit Binary Multiplication

Purpose: Multiply the 8-bit unsigned number In memory location 0040 by the 8-bit
unsigned number In memory location 0041. Place the eight least significant
bits of the result Into memory location 0042 and the eight most Significant
bits into memory location 0043.
Sample Problems:
a.

Result:

(0040)
(0041)

03
05

(0042)
(0043)

OF
00

or In deCimal 3 x 5 = 15
~

~04m

(0041)
Result:

6F
61

(0042)
OF
(0043)
2A
or 111 x 97 = 10.767

You can perform multiplication on a computer In the same way that you do long
multiplication by hand. Since the numbers are binary. the only problem IS whether to
multiply by 0 or 1; multiplying by zero obViously gives zero as a result. while multiplying
by one produces the same number that you started with (the multiplicand). So, each
step In a binary multiplication can be reduced to the following operation.
If the current bit in the multiplier IS 1. add the multiplicand
to the partial product.

-------.
MULTIPLICATION
ALGORITHM

The only remaining problem IS to ensure that you line everything up correctly each
lime. The following operations perform thiS task.
1)
2)

Shift multiplier left one bit so that the bit to be examined is placed Into the Carry.
Shift product left one bit so that the next addition is lined up correctly.

The complete process for binary mu Itiplicatlon is as follows: 1

Step 1 - InitializatIOn
Product = 0
Counter = 8
Step 2 - Shift Product so as to line up properly
Product = 2 x Product (LSB = 0)
Step 3 - Shift Multiplier so bit goes to Carry
Multiplier = 2 x MUltiplier

8-8

Step 4 - Add Multiplicand to Product if Carry IS 1

If Carry = 1. Product = Product + Multiplicand
Step 5 - Decrement Counter and check for zero
Counter = Counter - 1
If Counter =I 0 go to Step 2
In the case of Sample Problem b. where the multiplier is 61 (hex) and the multiplicand is
6F (hex) the process works as follows:
Initialization:
Product
Multiplier
Multiplicand
Counter

0000
61
6F
08

After first Iteration of steps 2-5:

Product
Multiplier
Multiplicand
Counter
Carry from Multiplier

0000
C2
6F
07
0

After second Iteration:

Product
Multiplier
Mu Itiplicand
Counter
Carry from Multiplier

006F
84
6F
06
1

After third iteration:

Product
Multiplier
Multiplicand
Counter
Carry from Multiplier

0140
08
6F
05
1

After fourth Iteration:

Product
Multiplier
Multiplicand
Counter
Carry from Multiplier

029A
10
6F
04
0

After fifth Iteration:

Product
Multiplier
Multiplicand
Counter
Carry from Multiplier

0534
20
6F
03
0

After sixth iteration:

Product
Multiplier
Multiplicand
Counter
Carry from Multiplier

OA68
40
6F
02
0

8-9

After seventh iteration:

Product
Multiplier
Multiplicand
Counter
Carry from Multiplier

1400
80
6F

01

After eighth Iteration:

Product
Multiplier
Multiplicand
Counter
Carry from Multiplier

2AOF
00
6F

00
1

Flowchart:

Multiplicand
Multiplier
Product
Count

(40l
(41l
0
8

Product = 2 x Product

(Shift left 1 bill

Multiplier =2xMultlplier

IShift left 1 bill

Product =
Product +

Multiplicand

Count == Count - 1

No

142 and 43l =

Product

8-10

Source Program:

MULT

CHCNT

LD
LD
LD
INC
LD
LD
LD
ADD
RLA
JR
ADD
DJNZ
LD
HALT

HLAOH
E.(HL)
0.0
HL
A.(HL)
HL.O
B.8
HL.HL

:GET MULTIPLICAND
:EXTEND TO 16 BITS

NC.CHCNT
HL.DE
MULT
(42HLHL

:GET MULTIPLIER
:PRODUCT = ZERO
:COUNT = BIT LENGTH OF MULTIPLIER
:SHIFT PRODUCT LEFT 1 BIT
;SHIFT MULTIPLIER LEFT 1 BIT
:IS CARRY FROM MULTIPLIER 17
:YES. ADD MULTIPLICAND TO PRODUCT
;SAVE PRODUCT IN MEMORY

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017

21
40
00
5E
16
00
23
7E
21
00
00
06
08
29
17
30
01
19
10
F9
22
42
00
76

Instruction
(Mnemonic)

MULT

CHCNT

LD

HL.40H

LD
LD

E.(HL)
0.0

INC
LD
LD

HL
A.(HL)
HL.O

LD

B.8

ADD
RLA
JR

HL.HL

ADD
DJNZ

HL.DE
MULT

LD

(42HI.HL

NC.CHCNT

HALT

Note that the multiplicand must be extended to 16 bits by clearing Register 0 so that It
can be added to the product using the ADD HL.DE instruction.
The Instruction ADD HUHL acts as a 16-blt logical left shift for the 16-blt product.
In this program. the Z80 16-blt instructions handle data rather than addresses. LD HL.O
is used to initialize the product; ADD HL.HL to perform a 16-bit logical left shift: ADD
HL.DE to add the multiplicand to the partial product; and LD (42HI.HL to store the
result in memory. You must be careful to extend 8-blt quantities (like the multiplicand
In this example) to 16 bits. Note that you cannot use the 16-blt facililles simultaneously
for addreSSing and data manipulation. However. if you have no other'need for the alternate registers. you could save the old contents of the regular registers there and restore

8-11

them afterward using the EXX instruction. This instruction exchanges the contents of
Register Pairs BC. DE. and HL with the contents of their alternate counterparts in just
four clock cycles.
Besides its obvious use in calculators and point-of-sale terminals. multiplication is a key
part of almost all signal processing and control algorithms. The speed at which
multiplications can be performed determines the usefulness of a CPU in process control. signal detection. and signal analysis.
The algonthm takes between 390 and 440 clock cycles to multiply on a Z80
microprocessor. The precise time depends on the number of one bits in the mUltiplier.
Other algorithms may be able to reduce the average execution time somewhat but 400
clock cycles will still be a typical execution time for a software muItiplication 2

8-Bit Binary Division

Purpose: Divide the 16-bit unSigned number in memory locations 0040 and 0041
(most significant bits In 0041) by the 8-blt unsigned number In memory location 0042. The numbers are normalized so that 1) the most significant bits of
both the dividend and the diVisor are zero and 2) the number in memory
location 0042 IS greater than the number in memory location 0041. i.e.. the
quotient is an 8-blt number. Store the quotient in memory location 0043 and
the remainder in location 0044.
Sample Problems:

a.

b.

(0040)
(0041)
(0042)

40 (64 decimal)
00
08

Result

(0043) = 08
(0044) =00
I.e.. 64/8 = 8

(0040)
(0041)
(0042)

6D (12.909 decimal!
32
47 (71 decimal!

Result

(0043) = B5 (181 decimal)

(0044) = 3A (58 decimal!
I.e.. 12.909/71 = 181 with a remainder of 58

You can perform division on the computer lust like you would perform diVISion with pen and paper. i.e.. using trial subtractions.
Since the numbers are binary. the only question IS whether the bit
In the quotient IS 0 or 1. I.e.. whether or not the divisor can be subtracted from what is
left of the dividend. Each step in a binary division can be reduced to the following
operation:
If the diVisor can be SUbtracted from the eight
most Significant bits of the diVidend Without
a borrow. the corresponding bit in the quotient IS 1. otherwise it IS O.
The only remaining problem is to line up the diVidend and quotient properly. You can
do this by shifting the dividend and quotient logically left one bit before each tnal
subtraction. The diVidend and quotient can share a 16-bit register. since the procedure
clears one bit of the diVidend at the same time as It determines one bit of the quotient.

8-12

The complete process for binary division is:

Step 1 - Initialization:
Quotient = 0
Counter = 8
Step 2 - Shift Dividend and Quotient so as to line up properly:
Dividend = 2 x Quotient
Quotient = 2 x Quotient
Step 3 - Perform trial Subtraction. If no Borrow add 1 to Quotient:
If 8 MSBs of Dividend> Divisor then
MSBs of Dividend = MSBs of Dividend - Divisor
Quotient = Quotient + 1
Step 4 - Decrement counter and check for zero:
Counter = Counter - 1
if Cou nter ;b O. go to Step 2
Remainder = 8 MSBs of Dividend
In the case of sample problem b. where the dividend is 3260 (hex) and the divisor IS 47
(hex). the process works as follows:
Initialization:
DiVidend
Divisor
Quotient
Counter

3260
47
00
00

After first iteration of Steps 2 - 4:

(Note that the diVidend is shifted prior to the trial subtraction)
Dividend 1DDA
Divisor
47
Quotient
01
Counter
07
After second iteration of Steps 2 - 4:
Dividend 3BB4
DiVisor
47
Quotient
02
Counter
06
After third Iteration:
DiVidend
DiVisor
Quotient
Counter

3068
47
05
05

Dividend
Divisor
Quotient
Counter

1900
47
OB
04

Dividend
Divisor
Quotient
Counter

33AO
47
16
03

After fourth iteration:

After fifth Iteration:

8-13

After sixth iteration:

Dividend
Divisor
Quotient
Counter

2040
47
2D
02

Dividend
Divisor
Quotient
Counter

4080
47
5A
01

Dividend
Divisor
Quotient
Counter

3AOO
47
B5
00

After seventh iteration:

After eighth Iteration:

So the quotient IS B5 and the remainder IS 3A.

The MSBs of dividend and divisor are assumed to be zero so as to simplify calculations
(the shift prior to the trial subtraction would otherwise place the MSB of the dividend In
the Carry). Problems that are not In this form must be simplified by removing parts of
the quotient that would overflow an 8-blt word. For example:
1024 = 400 (Hex) = 100 + 100 (Hex)
3 3 3
The last problem IS now

In

the proper form. An extra divIsion may be necessary.

8-14

Flowchart:

Dividend

(40 and 411

(42)
8
Quotient
= 0

Divisor
Count

8 MSBs of
Dividend = 8 MSBs
of Dividend - Divisor
Quotient =QuQtient + 1

Count

= Count - 1

No

(43l = Quotient
(44l = 8 MSBs of

Dividend

Source Program:

DIV:

CNT

LD
LD
LD
LD
ADD
LD
SUB
JR
LD
INC
DJNZ
LD
HALT

HL.(40Hl
A,(42Hl
C.A
B.8
HL.HL
A.H

;GET DIVIDEND
:GET DIVISOR
;COUNT = NUMBER OF BITS IN DIVISOR
;SHIFT DIVIDEND. QUOTIENT LEFT 1 BIT
;CAN DIVISOR BE SUBTRACTED?

C
C.CNT
H,A
L
DIV
(43Hl.HL

:NO. GO TO NEXT STEP

:YES, SUBTRACT DIVISOR FROM DIVIDEND
:ADD 1 TO QUOTIENT
;SAVE QUOTIENT. REMAINDER IN MEMORY

8-15

~~---~._---

Object Program:

Memorv Address
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015

2A
40
00
3A
42
00
4F
06
08
29
7C
91
38
02
67
2C
10
F7
22
43
00
76

Instruction
(Mnemonic)

DIV:

CNT:

LD

HL.(40H)

LD

A,(42H)

LD
LD

C,A
B,8

ADD
LD
SUB
JR

HL,HL
A,H
C
C,CNT

LD
INC
DJNZ

H,A
L
DIV

LD

(43Hl,HL

HALT

Register Pair HL holds both the dividend and the quotient. The quotient simply replaces
the dividend in Register L as the dividend is shifted left logically.
For longer division problems, you could use the instruction SBC HL which subtracts the
contents of a register pair and the contents of the Carry from the contents of Register
Pair HL.
The instruction INC L sets the least significant bit of the quotient to 1. since ADD HL,HL
has previously cleared that bit.
Division is necessary In calculators, terminals, communications error checking, control
algorithms, and many other applications.
This algorithm takes between 400 and 430 clock cycles to divide on a Z80
microprocessor. The precise time depends on the number of one bits in the quotient.
Other algorithms may reduce the average execution time somewhat. but 400 clock cycles will still be typical for a software divIsion. Some of the references listed at the end
of this chapter discuss faster methods for implementing division.

8-16

Self-Checking Numbers
Double Add Double Mod 10
Purpose: Calculate a checksum digit from a string of BCD digits. The length of the
string of digits (number of words) is in memory location 0041, the string of
digits (2 BCD digits to a word) starts in memory location 0042. Calculate the
checksum digit by the Double Add Double Mod 10 technique 3 and store it in
memory location 0040.

The Double Add Double Mod 10 technique works as follows:

1)
2)
3)
4)
5)

SELF-CHECKING
NUMBERS
Clear the checksum to start.
Multiply the leading digit by two and add the result to the
checksum.
Add the next digit to the checksum.
Continue the alternating process until you have used all the digits.
The least significant digit of the checksum IS the self-checking digit.

Self-checking digits are commonlv added to Identification numbers on credit cards. inventory tags, luggage, parcels. etc.. when they are handled by computerized systems.
They may also be used in routing messages. identifying files, and other applications.
The purpose of the digits is to minimize entry errors such as transposing digits (69 instead of 961. shifting digits (7260 Instead of 37261. missing digits by one (65 instead of
641. etc. You can check the self-checking number automatically for correctness upon
entry and can eliminate many errors immediately.
The analysis of self-checking methods is qUite complex. For example, a plain checksum
will not find transposition errors (4 + 9 = 9 + 4). The Double Add Double algOrithm will
find Simple transpOSition errors (2 x 4 + 9 = 17 '" 2 x 9 + 4); but will miss some errors,
such as transpositions across even numbers of digits (367 Instead of 763). However.
thiS method will find many common errors! The value of a method depends on what errors it will detect and on the probability of particular errors in an application.
For example, if the string of digits IS
549321
the result will be:
Checksu m
Se!f-checklng digit

5 x 2 + 4 + 9 x 2 + 3 + 2 x 2 + 1 = 40
0 (least significant digit of a checksum)

Note that an erroneous entry like 543921 would produce a different self-checking digit
(41. but erroneous entries like 049321 or 945321 would not be detected.
Sample Problems:

a.

(0041)
(0042)
(0043)
(0044)

03
36
68
51

Result: Checksum = 3 x 2 + 6 + 6 x 2
(0040)
03
b.

(0041)
(0042)
(0043)
(0044)
(0045)

+8 +5 x 2 + 1 =

43

04
50
29
16
B3

Result: Checksum = 5 x 2 + 0 + 2 x 2
(0040) = 00

8-17

+9+ 1x 2+6+8 x2+3 =

50

Flowchart:

Checksum
Count
Pointer

POinter

0
(41)
42

POinter + 1

Count = Count - 1

(40) = Checksum

AND 000011116

8-18

Source Program:

CHDIG.

LD
LD
LD
LD
LD
LD
RRA
RRA
RRA
RRA
AND
ADD
DAA
ADD
DAA
LD
LD
AND
ADD
DAA
LD
INC
DJNZ
AND
LD
HALT

A, (41 H)
B.A
C.O
HL,42H
A.(HU
D.A

00001111B
A.A
A.C
C,A
A.D
00001111B
A.C
C.A
HL
CHDIG
00001111B
(40Hl.A

;COUNT = LENGTH OF STRING IN BYTES

;CHECKSUM = 0
;POINT TO START OF STRING OF DIGITS
;GET TWO BCD DIGITS FROM STRING
;SAVE COpy
;GET MSD BY SHIFTING AND MASKING

;DOUBLE MSD
;MAKE DOUBLED MSD DECIMAL
;ADD DOUBLED MSD TO CHECKSUM
;KEEP CHECKSUM DECIMAL
;GET LEAST SIGNIFICANT DIGIT
;(MASK OUT MSD)
;ADD LSD TO CHECKSUM
;KEEP CHECKSUM DECIMAL

;MASK OFF SELF-CHECKING DIGIT

;SAVE SELF-CHECKING DIGIT

8-19

Object Program'
Memory Address
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
0008
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A
0018
001C
0010
001E
001F
0020
0021
0022
0023
0024

3A
41
00
47
OE
00
21
42
00
7E
57
1F
1F
1F
1F
E6
OF
87
27
81
27
4F
7A
E6
OF
81
27
4F
23
10
EA
E6
OF
32
40
00
76

Instruction
(Mnemonic)

CHDIG:

LD

A.(41H)

LD
LD

8.A
C.O

LD

HL.42H

LD
LD
RRA
RRA
RRA
RRA
AND

A(HU
D.A

ADD
DAA
ADD
DAA
LD
LD
AND

000011118
A.A
AC
C.A
A.D
000011118

ADD
DAA

A.C

LD

C.A

INC
DJNZ

HL
CHDIG

AND

000011118

LD

(40H),A

HALT

The digits are removed by shifting and masking. Four right shifts are needed to separate
out the most significant digit.
A decimal adjust (DAA) must follow each addition to produce the proper decimal result.
A single DAA after a series of additions will not work (try ItO. Remember that DAA
works only on the Accumulator.
There IS no problem with carries from the deCimal sum. since the procedure uses only
the least Significant digit of the checksum anyway.

8-20

DECIMAL SHIFT
An alternative (and superior) approach IS to use the Z80
decimal shift instruction RLD. This instruction IS a 4-blt
INSTRUCTIONS
shift that moves the contents of the four least significant
bits of the memory location addressed by HL Into the four most significant bits of that
location. the previous contents of the four most significant bits of that location Into the
four least significant bits of the Accumulator. and the previous contents of the four
least significant bits of the Accumulator Into the four least significant bits of the memory location. Thus. RLD not only moves a single digit to the Accumulator. but it also
shifts the next digit so that it can be moved to the Accumulator with the next RLD.
Figure 8-1 shows an example of how RLD works; RRD is the same instruction except
that the shift is right instead of left.

The Double Add Double Mod 10 algorithm can be implemented as follows uSing RLD:
Source Program:

CHDIG;

LD
LD
LD
LD
SUB
RLD
ADD
DAA
ADD
DAA
LD
SUB
RLD
ADD
DAA
LD
INC
DJNZ
AND
LD
HALT

A.(41Hl
B.A
C.O
HL,42H
A
A.A
A.C
C.A
A
A.C
C.A
HL
CHDIG
00001111B
(40H),A

;COUNT =LENGTH OF STRINGS (IN BYTES)

;CHECKSUM = 0
;POINT TO START OF STRING OF DIGITS
;CLEAR MSD
;GET MSD FROM STRING
;DOUBLE MSD
;MAKE DOUBLED MSD DECIMAL
;ADD DOUBLED MSD TO CHECKSUM
;KEEP CHECKSUM DECIMAL
;CLEAR MSD
;GET LSD FROM STRING
;ADD LSD TO CHECKSUM
;KEEP CHECKSUM DECIMAL

;MASK OFF SELF-CHECKING DIGIT

;SAVE SELF-CHECKING DIGIT

8-21

Object Program:
Memorv Address
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB.
OOOC
OOOD
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A
001B
001C
0010
001E
001F

3A
41
00
47
OE
00
21
42
00
97
ED
6F
87
27
81
27
4F
97
ED
6F
81
27
4F
23
10
EF
E6
OF
32
40
00
76

Instruction
(Mnemonic)

CHDIG:

LD

A.(41H)

LD
LD

B.A
C.O

LD

HL42H

SUB
RLD

ADD
DAA
ADD
DAA
LD

A.A

C.A

SUB

A.C

RLD
A,C

ADD
DAA
LD
INC
DJNZ

C.A
HL
CHDIG

AND

00001111B

LD

(40H),A

HALT

We could Improve this program even further (it is already shorter than the prevIous version), Since we are dropping the most significant digit at the end anyway. there is no
reason to clear it out each time with the SUB A instruction,

8-22

Initial Conditions
{HLI =4000
{AI =7F
{4oool = 12
After ALD

{AI =71
{40001 =2F
Before

After

Accumulator

Memorv

Rit2

,.. . . . "

fF< [E]

~/I
',

High-order bit = 0,

~::~:

Memory

Accumuiator

I I

: ..
: .." 1-

Non-zero result,

4 ones, set PIG to 1

set2toO

After AAD

{AI =72
{4oool = F
Before

After
Memory

Accumulator

Accumulator

- -

Memorv

rn

High-order bit = O. set S to 0 ....I---------...J

4 ones, set P/0 to 1

Figure 8-1. Examples of the Z80 Digit Shifts

You can double a decimal number (in the Accumulatod by
adding It to Itself and then performing a decimal correction,
I.e.,

ADD
DAA

;DOU8LE NUM8ER
;AND MAKE RESULT DECIMAL

DOUBLING
AND HALVING
BINARY
NUMBERS

Remember that the Accumulator can hold only valid decimal digits in the range 0~99.
You cannot use SLA A (Shift Left ArithmetiC AI because that instruction always clears
the Half-Carry (only Add and Subtract instructions set H properly).
You can divide a decimal number by two simply by shifling it right logically and then
subtracting three from any digit that is eight or larger (since 10 BCD IS 16 binary). The
following program divides a decimal number In memory location 0040 by two and
places the result Into memory location 0041.

8-23

DONE:

LD
SRL
BIT
JR
SUB
LD
HALT

A. (40Hl

A
3,A
Z.DONE

3
(41H),A

GET DECIMAL NUMBER

DIVIDE BY 2 IN BINARY
IS LEAST SIGNIFICANT DIGIT 8 OR MORE?
:YES. SUBTRACT 3 FOR DECIMAL CORRECTION
:STORE NUMBER DIVIDED BY 2

Try this program and the method on the decimal numbers 28, 30, and 37 Do you understand why It works?
Rounding is simple whether the numbers are binary or decimal. A
binary number can be rounded as follows:
If the most significant bit to be dropped IS 1,
add 1 to the remaining bits. Otherwise. leave
the remaining bits alone.
This rule works because 1 is halfway between 0 and 10 in binary, much as 5 IS halfway
in decimal (note that 0.5 decimal = 0.1 binary).
So, the following program will round a 16-blt number in memory locations 0040 and
0041 (MSBs in 0041) to an 8-bit number in memory location 0041.

DONE,

LD
BIT
JR
INC
INC
HALT

HLAOH
7,(HU
Z.DONE
HL
(HU

:IS MSB OF EXTRA BYTE 1?

;NO. ROUND UP

If the number IS longer than 16 bits. the rounding must npple through the other bytes
as needed.
Decimal rounding IS a bit more difficult because the crossover
POint IS now BCD 50 and the rounding must produce a decimal
result. The rule IS:
If the most Significant digit IS to be dropped
IS 5 or more, add 1 to the remaining digits.
The following program will round a 4-digit BCD number in memory locatiOns 0040 and
0041 (MSBs in 0041) to a 2-diglt BCD number in memory locatiOn 0041.

DONE,

LD
LD
CP
JR
INC
LD
ADD
DAA
LD
HALT

HL.40H
A.(HU
50H
C,DONE
HL
A.(HU
A.1

:IS BYTE TO BE DROPPED 50 OR MORE?

:YES. ROUND MSB'S UP

:KEEP DIGITS DECIMAL

(HU,A

Remember that the DAA instruction works only on numbers in the Accumulator. In this
case, we could round with the instruction INC A. since we know that the Carry IS zero
(Why? - remember the JR instruction). Normally, we need the sequence ADD A,l
followed by DAA. since INC A does not affect the Carry.

8-24

SIGN
Very often when performing multibyte twos complement
signed arithmetic. it is necessary to propagate the sign bit
PROPAGATION
through the high-order bytes. This operation can be performed
in a straightforward manner if. as is usually the case. the sign is in the Carry. The SSC
A.A Instruction has the effect of propagating the state of the Carry throughout a word.
Since A-A always equals O. SSC A.A IS equivalent to subtracting the Carry from 0 and
can Yield only the values 0 and FFH.

PROBLEMS
1) Multiple-Precision Subtraction
Purpose: Subtract one multiple-word number from another. The length of the numbers is In memory location 0040. the numbers themselves start (least significant bits first) in memory locations 0041 and 0051. respectively. and the
difference replaces the number starting in memory location 0041. Subtract
the number starting in 0051 from the one starting In 0041.
Sample Problem:

Result:

that

2)

(0040)

04

(0041)
(0042)
(00431
(00441

C3
A7
58
2F

(0051)
(00521
(0053)
(0054)

88
35
DF
14

(0041)
(0042)
(0043)
(0044)

08
72
7C
1A
2F5SA7C3
14DF35S8
1A7C7208

IS.

Decimal Subtraction

Purpose: Subtract one multiple-word decimal (SCDI number from another. The length
of the numbers IS in memory location 0040. the numbers themselves start
(least significant bits first) in memory locations 0041 and 0051. respectively.
and the difference replaces the number starting in memory location 0041.
Subtract the number starting In 0051 from the one starting In 0041.
Sample Problem:

(0040)

04

(0041)
(0042)
(0043)
(00441

85
19
70
36

(0051)
(0052)
(0053)
(0054)

59
34
66
12

8-25

Result:

(0041)
(0042)
(0043)
(0044)

26
85
03
24
36701985
12663459

that IS,

24038526

3)

8-Bit by 16-Bit Binary Multiplication

Purpose: Mu Itlply the 16-blt unsIgned number in memory locations 0040 and 0041
(most significant bits In 0041) by the 8-blt unsigned number In memory location 0042. Store the result In memory locations 0043 through 0045, with the
most Significant bits In memory location 0045.
Sample Problems:
a.

Result:

(0040)
(0041)
(0042)

03
00
05

(0043)
(0044)
(0045)

OF
00
00
3 x 5 = 15

that IS,
b.

Result:

(0040)
(0041)
(0042)

6F
72 (29,295 decimal!
61 (97 decimal!

(0043)
(0044)
(0045)

OF

29.295 x 97

that IS,

4)

5C

28
2,841.615

Signed Binary Division

Purpose: Divide the 16-blt signed number in memory locations 0040 and 0041 (most
Significant bits in 0041 by the 8-blt signed number In memory location 0042.
The numbers are normalized so that the magnitude of memory location 0042
IS greater than the magnitude of memory location 0041. Store the quotient
(Signed) In memory location 0043 and the remainder (always positive) In
memory location 0044.
Sample Problems:
a.

Result:

Result:

co

(0040)
(0041)
(0042)

FF (-64)
08

(0043)
(0044)

F8 (-8) quotient
00 (0) remainder

(0040)
(0041)
(0042)

ED (-4717)
47 (71 deCimal!

(0043)
(0044)

SO (-67 deCimal!
28 (+40 deCimal!

93

Hint: Determine the sign of the result. perform an unsigned diVision, and adlust the quotient and remainder properly.

8-26

5)

Self-Checking Numbers Aligned 1, 3, 7 Mod 10

Purpose: Calculate a checksum digit from a string of BCD digits. The length of the
string of digits (number of words) IS In memory location 0041. the string of
digits (2 BCD digits to a word) starts in memory location 0042. Calculate the
checksum digit by the Aligned 1. 3. 7 Mod 10 method and store It In memory
location 0040.
The Aligned 1. 3. 7 Mod 10 technique works as follows:
1)

Clear the checksum to start.

2)

Add the leading digit to the checksum.

3)

Multiply the next digit by 3 and add the result to the checksum.

41
5)
6)

Multiply the next digit by 7 and add the result to the checksum.
Continue the process (Steps 2-4) until you have used all the digits.
The self-checking digit IS the least significant digit of the checksum.

For example. if the string of digits IS:

549321
the result will be:
Checksum
Self-checking digit

5 + 3 x 4 + 7 x 9 + 3 + 3 x 2 + 7 x 1 = 96

Sample Problems:
a.

(0041)
(0042)
(0043)
(0044)

03
36
68
51

Result: Checksum = 3 + 3 x 6 + 7 x 6 + 8 + 3 x 5 + 7 x 1 = 93
03
(0040)
b.

(0041)
(0042)
(0043)
(0044)
(0045)

04
50
29
16
83

Result: Checksum = 5 + 3 x 0 + 7 x 2 + 9 + 3 x 1 + 7 x 6 + 8
+3x3=90
(0040) = 00

Hint: Note that 7 2 x 3 + 1 and 3 2 x 1 + 1 so the formula

Mi = 2 x Mi-1 + 1 can be used to calculate the next multiplYing factor.

8-27

REFERENCES
1.

Several multiplication algorithms are described in T. Dollhoff. "Microprocessor

Software: How to Optimize Timing and Memory Usage. Part Four. Techniques for
the Zilog Z80." Digital Design. February 1977. pp. 44-51.

2.

Some microprocessors (such as the 9900. 8086. and Z-8000) have hardware
multiplication Instructions that are somewhat faster. but maximum speed requires
the addition of external hardware.
Other methods for implementing multiplication. divIsion. and other arithmetic tasks
are discussed In:
Geist. D. J.. "MOS Processor Picks up Speed with Bipolar Multipliers," Electronics.
Ju Iy 7. 1977, pp. 113-115.
Kolodzlnski. A. and D. Wainland. "MultiplYing with a Microcomputer," Electronic
Design. January 18. 1978. pp. 78-83.
Mick. J. R. and J. Springer. "Single-chip MUltiplier Expands Digital Role in Signal
Processing," Electronics. May 13. 1976. pp. 103-108.
Parasuraman. B.. "Hardware Multiplication Techniques for Microprocessor
Systems." Computer Design. April 1977, pp. 75-82.
Tao. 1. F. et al.. "Applications of Microprocessors in Control Problems," 1977 Joint
Automatic Control Conference Proceedings. San Francisco. CA" June 22-24. 1977
Waser. S.. "State-of-the-art in High-Speed Arithmetic Integrated Circuits," Computer Design. July 1978. pp. 67-75.
-Weissberger. A. J. and T. Toal. "Tough Mathematical Tasks Are Child's Play for
Number Cruncher," Electronics. February 17, 1977. pp. 102-107

3.

See J. R. Herr. "Self-Checking Number Systems," Computer Design. June 1974.

pp.85-91.

8-28

Chapter 9
TABLES AND LISTS
Tables and lists are two of the basic data structures used with all computers. We have
already seen tables used to perform code conversions and anthmetlc. Tables may also
be used to identify or respond to commands and instructions. lineanze data. provide access to files or records. define the meaning of keys or sWitches. and choose among
alternate programs. lists are usually less structured than tables. lists may record tasks
that the processor must perform. messages or data that the processor must record. or
conditions that have changed or should be monitored. Tables are a simple way of making decisions or solving problems. since no computations or logical functions are
necessary. The task. then. reduces to organizing the table so that the proper entry IS
easy to find. lists allow the execution of sequences of tasks. the preparation of sets of
results. and the construction of interrelated data files (or data bases), Problems Include
how to add elements to a list and remove elements from it.

EXAMPLES
Add Entry to List
Purpose:

Add the contents of memory location 0040 to a list if It IS not already present in the list. The length of the list is In memory location 0041 and the list
Itself begins In memory location 0042.

Sample Problems:

Resu It:

~04m

68

(0041)
(0042)
(0043)
(0044)
(0045)

04
37
61
38
1D

(0041)
(0046)

05
68

The entry is added to the list. since it IS not already present. The length of the list IS increased by 1.

b.

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)
Result: No change.

68
04
37
68
38
lD
since the entry IS already In the list.

9-1

Flowchart:

Entry
Count

(40)
(411

Pointer

42

Pointer
Count

= POinter

+ 1

= Count - 1

(Pointed
(411

=
~

Entry
(411 + 1

Source Program:

SRLST

DONE:

LD
LD
INC
LD
INC
CP
JR
INC
DJNZ
LD
LD
INC
HALT

HLAOH
A,(HU
HL
B,(HU
HL
(HU
Z.DONE
HL
SRLST
(HU,A
HLA1H
{HU

:POINT TO ENTRY
:GET ENTRY
:POINT TO COUNT
:COUNT = LENGTH OF LIST
:POINT TO START OF LIST
:IS ENTRY = ELEMENT IN LIST?
:YES. THROUGH
:NO. GO ON TO NEXT ELEMENT
:ADD ENTRY TO LIST
:ADD 1 TO LIST LENGTH

9-2

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012

21
40
00
7E
23
46
23
BE
28
08
23
10

Instruction
(Mnemonic)

SRLST

LD

HLAOH

LD
INC
LD
INC
CP
JR

A.(HL)
HL
B.(HL)
HL
(HL)
Z.DONE

INC
DJNZ

HL
SRLST

LD
LD

(HL).A
HLA1H

INC
HALT

(HL)

FA
77
21
41
00
34
76

ADELM:

DONE.

We could also use the block search instruction CPIR

In

our example. as follows:

Source Program:

DONE:

LD
LD
INC
LD
LD
INC
CPIR
JR
LD
LD
HALT

HL.40H
A.(HU
HL
B.O
C.(HL)
HL
Z.DONE
(HL).A
HLAl H

;POINT TO ENTRY
:GET ENTRY
;POINT TO COUNT
;COUNT = LENGTH OF LIST (16 BITS)
:POINT TO START OF LIST
:LOOK FOR ENTRY IN LIST
:DONE IF ENTRY FOUND
;OTHERWISE. ADD ENTRY TO LIST
:ADD 1 TO LIST LENGTH

9-3

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012

40
00
7E
23
06
00
4E
23
ED
Bl
28
05
77
21
41
00
34
76

Instruction
(Mnemonic)

21

DONE:

LD

HLAOH

LD
INC
LD

A.(HL)
HL
B.O

LD
INC
CPIR

C.(HL)
HL

JR

Z.DONE

LD
LD

(HL).A
HL,41 H

INC
HALT

(HU

Remember that CPIR automatically repeats the basic Search instruction until either BC
is decremented to zero or a true comparison occurs (i.e.. A = (HUl.
Be careful of the following slight differences from the prevIous version:
1)
2)

BC is a 16-blt counter. Thus. CPIR can handle strings longer than 256 bytes.
The Panty/Overflow bit (P/0) IS cleared if BC is decremented to zero. and set otherwise.

Clearly. thiS method of adding elements IS very ineffiCient if the list

HASHING
is long. We could improve the procedure by limiting the search to
part of the list or by ordering the list. We could limit the search by using the entry to get
a starting POint In the list. This method IS called "hashing". and IS much like selecting a
starting page In a dictionary or directory on the basIs of the first letter In an entry. We
could order the list by numerical value. The search could then end when the list values
went bevond the entry (larger or smaller. depending on the ordenng technique used). A
new entry would have to be Inserted properly. and all the other entnes would have to be
moved down In the list.
The program could be restructured to use two tables. One table could provide a starting
POint in the other table; for example. the search point could be based on the most or
least significant 4-bit digit in the entry.

9-4

The program does not work if the length of the list could be zero (what happens?). We
cou Id avoid this problem by checking the length Initially. The initialization procedure
for the first program would then be:

ADELM:

LD
LD
INC
LD
INC
INC
DEC
JR

HLAOH
A.(HL)
HL
8.(HL)
HL
8
8
Z.ADELM

:POINT TO ENTRY
:GET ENTRY
:POINT TO LENGTH
:COUNT = LENGTH OF LIST
:POINT TO START OF LIST
:IS COUNT ZERO?

LD

{HL),A

:ADD ENTRY TO LIST

:YES. GO ADD ENTRY TO LIST

Note that the sequence INC. DEC IS an easy way to check for a zero value In a register
without using the Accumulator or changing the value in the register.
The procedure:
LD
INC

HL.ADDR
(HU

IS a quick way to add 1 to a counter in memory location ADDR without uSing the Accumulator. You can use DEC (HL) in a similar manner to subtract 1 from the counter. LD
(HL),CONST can place a starting value (such as zero) In the counter. Memory locations
should. of course. be used for counters only when no readily accessible registers are
available.
If each entry were longer than one word. a pattern-matching program would be necessary. The program would have to proceed to the next entry if a match failed: that is. skip
over the last part of the current entry once a mis-match was found.

Check an Ordered List

Purpose:

Check the contents of memory location 0041 to see if It IS in an ordered list.

The length of the list IS in memory location 0042: the list itself begins In
memory location 0043 and consists of unsigned binary numbers in increasIng order. If the contents of location 0041 is in the list. clear memory location 0040: otherwise. set memory location 0040 to FF (hex).

Sample Problems:

a.

Resu It:
b.

(0041)
(0042)
(0043)
(0044)
(0045)
(0046)

68
04

(0040)

FF. since 68 IS not in the list.

68
04

(0041)
(0042)
(0043)
(0044)
(0045)
(0046)
Result:

(0040)

37
55
7D
A1

37
55
68
A1
00. since 68 IS In the list.

9-5

Flowchart:

Entry
Count
Pointer
Mark

Pointer
Count

=
=

(4t)
(42)
43
0

+'

Pointer
Count - 1

Mark = FF (Hax)

(40) = Mark

The searching process IS a bit different here since the elements are ordered. Once we
find an element larger than the entry. the search is over, since subsequent elements will
be even larger. You may want to try an example to convince yourself that the procedure
works.
As In the prevIous problem. a table or other method that could
choose a good starting point would speed up the search. One
method would be to start in the middle and determme which half
of the list the entry was In. then divide the half into halves. etc. This method IS called a
binary search. since it divides the remaining part of the list In half each time.'

9-6

Source Program:

SRLST"

NOTIN:
DONE,

LD
LD
INC
LD
LD
INC
CP
JR
JR
INC
DJNZ
LD
LD
LD
HALT

HL,41H
A,(HL)
HL
B,(HL)
C,O
HL
(HL)
Z,DONE
C,NOTIN
HL
SRLST
C.OFFH
A.C
(40Hl.A

;POINT TO ENTRY
:GET ENTRY
;POINT TO LENGTH
;COUNT = LENGTH OF LIST
;MARK = ZERO FOR IN LIST
;POINT TO START OF LIST
:IS ENTRY = ELEMENT IN L1sn
;YES SEARCH COMPLETED
;ENTRY NOT IN LIST IF LESS THAN ELEMENT

;MARK = FF FOR NOT IN LIST

;SAVE MARK

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD

21
41
00
7E
23
46
OE
00
23
BE
28
07
38
03
23
10
F8
OE
FF
79
32
40
00
76

OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017

Instruction
(Mnemonic)

SRLST:

NOTIN:

LD

HL.41 H

LD
INC
LD
LD

A,(HL)
HL
B.(HL)
C.O

INC
CP
JR

HL
(HL)
Z.DONE

JR

C,NOTIN

INC
DJNZ

HL
SRLST

LD

C.OFFH

LD
LD

A.C
(40Hl.A

HALT

The Z80 block search instructions are not as usefu I here as In the prevIous example
because we want to do more than a simple search. Now we also want to check to see if
we have examined the relevant part of the list (j.e.. the part where the elements are less
than or equal to the entry). Try rewriting the program to use CPl. Remember that you
must use the Parity/Overflow flag to determine if the byte counter has been decremented to zero.

9-7

Remove Element from Queue

Purpose:

Memory locations 0042 and 0043 contain the address of the head of the
queue (MSBs In 0043). Place the address of the first element (head) of a
queue into memory locations 0040 and 0041 (MSBs in 0041) and update
the queue to remove the element. Each element In the queue IS two bytes
long and contains the address of the next two-byte element in the queue.
The last element In the queue contains zero to Indicate that there IS no next
element.

Queues are used to store data In the order in which it will be used. or tasks in the order
In which they will be executed. The queue IS a first-in. first-out data structure: I.e.. elements are removed from the queue In the same order In which they were entered.
Operating systems place tasks in queues so that they will be executed in the proper
order. I/O drivers transfer data to or from queues so that it will be transmitted or
handled in the proper order. Buffers may be queued so that the next available one can
easily be found and those that are released can easily be added to the available storage.
Queues may also be used to link requests for storage. timing. or I/O so that they can be
satisfied In the correct order.
In real applications each element in the queue will typically contain a large amount of
information or storage space besides the address required to link the element to the
next one.
Sample Problems:
a.

Result:

b.
Result:

(0042)
(0043)
(0046)
(0047)
(004D)
(004E)

6~}

address of first element in queue

6~}

address of second element in queue

~~}

end of queue

(0040)
(0041)
(0042)
(0043)

~g}

address of element removed from queue

6~}

address of new first element in queue

(0042)
(0043)

~~}

empty queue

(0040)
(0041)

~~}

no element available from queue

9-8

Flowchart:

Pointer = (42 and 431

(40 and 411 = Pointer

(42)
{431

= (Pointad
= (Pointer + 1I

Source Program:

DONE:

LD
LD
LD
OR
JR
LD
INC
LD
LD
HALT

HL.(42H)
(40H).HL
A.H
L
Z.DONE
E.(HL)
HL
D.(HL)
{42H).DE

:GET ADDRESS OF HEAD OF QUEUE

:REMOVE HEAD OF QUEUE
:IS QUEUE EMPTY?
:YES. DONE
:NO. GET ADDRESS OF NEXT ELEMENT

:MOVE NEXT ELEMENT TO HEAD OF QUEUE

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011

2A
42
00
22
40
00
7C
B5
28
07
5E
23
56
ED
53
42
00
76

Instruction
(Mnemonic!

DONE:

9-9

LD

HL.(42H)

LD

(40Hl.HL

LD
OR
JR

A.H
L
Z.DONE

LD
INC
LD
LD

E.(HL)
HL
D.(HL)
(42H).DE

HALT

Queuing can handle lists that are not in sequential memory locations. Each element
must contain the address of the next element. Such lists allow You to handle data or
taskS In the proper order. change variables. or fill In definitions In a program. Extra
storage IS required. but elements can easily be added to the queue or deleted from It.
Note the use of the sequence:
LD
OR

A.H
L

to determine if the contents of a 16-blt register pair is zero. Remember that INC and
DEC do not affect any flags when applied to a register pair. Try to devise some other sequences that could handle this problem - I t obviously occurs whenever you use a 16bit counter rather than the 8-blt counter that we have used In most of the examples.
One problem is that there IS no instruction that loads a register pair uSing the address in
a register pair. A sequence of instructions is necessary whenever a register pair must be
loaded directly.
It may be useful to maintain pOinters to both ends of the queue rather than Just to its
head. The data structure may then be used in either a first-in. first-out manner or in a
last-In. first-out manner. depending on whether new elements are added to the head or
the tail. How would you change the program example so that memory locations 0044
and 0045 contain the address of the last element (tail) of the queue?
If there are no elements in the queue. the program clears memory locations 0040 and
0041. A program that requested an element from the queue would then have to check
those memory locations to see if its request had been satisfied. Can you suggest other
ways to provide this Information?

8-Bit Sort
Purpose:

Sort an array of unsigned binary numbers into descending order. The length
of the array is In memory location 0040 and the array Itself beginS In memory location 0041.

Sample Problem:

Resu It:

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)
(0046)

06
2A
B5
60
3F
D1
19

(0041)
(0042)
(0043)
(0044)
(0045)
(0046)

D1
B5
60
3F
2A
19

A simple sorting technique works as follows:

SIMPLE
SORTING
ALGORITHM

Step 1)

Clear a flag INTER.

Step 2)

Examine each consecutive pair of numbers in the array. If

any are out of order. exchange them and set INTER.
If INTER = 1 after the entire array has been examined. return to Step 1.

Step 3)

9-10

INTER will be set if any consecutive pair of numbers IS out of order. Therefore. if INTER = 0 at the end of a pass through the entire array. the array IS In proper order.
This sorting method is referred to as a "bubble sort" It IS an easy algorithm to implement. 'However. other sorting techniques should be considered when sorting long lists
where speed IS Important. 2
The technique operates as follows In a simple case. Let us assume that we want to sort
an array into descending order; the array has four elements - 12, 03, 15, 08.
1st Iteration;
Step 1)

INTER = 0

Step 2)

Final order of the array IS;

12
15
08
03
since the second pair (03.15) is exchanged and so is the third pair (03,08).
INTER = 1.

2nd Iteration;
Step 1)
Step 2)

INTER = 0
Final order of the array IS;
15
12
08
03
since the first pair (12,15) is exchanged. INTER = 1.

3rd Iteration;
Step 1)

INTER = 0

Step 2)

The elements are already in order. so no exchanges are necessary and INTER
remains zero.

9-11

Flowchart:

Inter

Count
POinter

o
(4Ol-1
41

Temp = (Pointer)
(Pointer) = (Pointer + 1)
IPointer+1l = Temp
Inter = 1

Pointer

= Pointer

+1

Count = Count - 1

No

No

9-12

Source Program:
SORT:

PASSl

CNT

LD
LD
LD
DEC
INC
LD
INC
CP
JR
LD
LD
DEC
LD
INC
LD
DJNZ
DEC
JR
HALT

C.O
HL,40H
B.(HU
B
HL
A,(HL)
HL
(HU
NC.CNT
D.(HL)
(HL).A
HL
(HU.D
HL
C.l
PASSl
C
Z.SORT

;CLEAR INTERCHANGE FLAG

;COUNT = LENGTH OF ARRAY
;NUMBER OF PAIRS = COUNT-1
;POINT TO START OF ARRAY
;GET ELEMENT FROM ARRAY
;IS IT LESS THAN NEXT ELEMENT?
;NO. NO INTERCHANGE NECESSARY
;YES. INTERCHANGE ELEMENTS

;SET INTERCHANGE FLAG

:WAS INTERCHANGE FLAG SET?
;YES. DO ANOTHER PASS

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
0008
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019

OE

Instruction
(Mnemonic)
SORT

00
21
40
00
46
05
23
7E
23
BE
30
07
56

PASSl

77
28

72
23
DE
01
10
F2
OD
28
E7
76

CNT

LD

C.O

LD

HLAOH

LD
DEC
INC
LD
INC
CP
JR

B.(HU
B
HL
A,(HU
HL
(HL)
NC.CNT

LD
LD
DEC
LD
INC
LD

D.(HL)
(HL).A
HL
(HL).D
HL
C.l

DJNZ

PASSl

DEC
JR

C
Z.SORT

HALT

The case where two elements In the array are equal is very Important here. The program
should not perform an interchange in that case. since that Interchange would occur in
every pass. The result would be that every pass would set the Interchange flag. thus
producing an endless loop.

9-13

The program must reduce the counter by 1. since the number of consecutive pairs I'
one less than the number of elements (the last element has no successor). Before start
ing each sorting pass, we must be careful to relnitialize the counter, pOinter. and in
terchange flag.
There are many possible minor variations on this program. For example, we could use
RES O,C and SET O,C to clear and set the interchange flag instead of LD C.O and LD C.1.
We could also use the sequence MOV B.C followed by DJNZ SORT to check the in
terchange flag.
Note that Register B should be used for the Inner counter, since that counter is decremented most frequently. This allows us to take maximum advantage of the DJNZ instruction.
Indexing would be a convenient way to perform the Interchange if the Z80's Index
registers were more accessible. Try rewriting the program so as to use one of the Index
registers and compare the execution time and memory usage of the rewritten program
to those of the Original program.

Using an Ordered Jump Table

Use the contents of memory location 0040 as an index to a jump table startIng In memory location 0041. Each entry In the Jump table contains a 16-bit
address with LSBs in the first word. The program should transfer control to
the address with the appropriate Index; that is. if the index is 6, the program lumps to address entry #6 In the table. Assume that the table has
fewer than 128 entries.

Purpose:

Sample Problem:

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)
(0046)
(0047)
(0048)
Resu It:

(PC)

02
48
00
4C
00
50
00
54
00
0050, since that is entry #2.
(starting from zero) In the Jump table.

Flowchart:

,
,

Start

Index
Base

=
=

(401 x Z
41

JELEM =
Base + Index

t
(PCl =
(JELEMl (JELEM + 11

9-14

The last box results in a transfer of control to the address obtained from the table.
Source Program:

LD
LD
ADD
LD
LD
INC
ADD
LD
INC
LD
EX
JP

HLAOH
A.(HL)
A.A
E.A
D.O
HL
HL.DE
E.(HL)
HL
D.(HL)
DE.HL
(HL)

;POINT TO INDEX
:GET INDEX
:DOUBLE INDEX FOR 2-BYTE TABLE
;EXTEND INDEX TO 16 BITS
;BASE ADDRESS OF JUMP TABLE
:INDEX INTO JUMP TABLE
:GET LSB'S OF DESTINATION ADDRESS
;GET MSB'S OF DESTINATION ADDRESS
:TRANSFER CONTROL TO DESTINATION

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE

21
40
00
7E
87
5F
16
00
23
19
5E
23
56
EB
E9

Instruction
(Mnemonic)
LD

HLAOH

LD
ADD
LD
LD

A.(HL)
AA
E.A
D.O

INC
ADD
LD
INC
LD
EX
JP

HL
HL.DE
E.(HL)
HL
D.(HL)
DE.HL
(HL)

Jump tables are very useful in situations where one of several routines must be
selected. Such situations arise in decoding commands. selecting test programs. choosing alternate methods. or selecting an I/O configu ration.
The lump table replaces a whole series of conditional Jump operations. The program
that accesses the jump table could be used to access several different tables merely by
changing the starting address. 3
The data must be multiplied by two to give the correct Index. since each entry in the
Jump table IS a two-byte address.
The instruction JP (HL). which transfers the contents of Register
Pair HL to the Program Counter. IS an indirect Jump that IS very
handy in Jump tables and monitor programs. Note that JP (HL) is a
Jump instruction. Since it places a new value into the Program Counter; however. it
allows us to place a vanable address directly into the Program Counter. All of the Conditional Jump instructions (and the Call instructions) use fixed addresses. The only Jump
Instructions with similar flexibility are the two-word instructions JP (IX) and JP (lY).
No ending operation IS necessary. Since JP (HL) transfers control to the address obtained from the Jump table.

9-15

PROBLEMS
1)

Remove an Entry From a List

Purpose:

Remove the contents of memory location 0040 from a list if It is present.

The length of the list is in memory location 0041 and the list itself begins In
memory location 0042. Move the entries below the one removed up one
POSitiOn and reduce the length of the list by 1.

Sample Problems:

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)

a.

68
04

37
61
28

10

Result: No change. since the entry is not in the list.

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)

b.

Resu It:

(0041)
(0042)
(0043)
(0044)

68
04

37
68
28
1D
03

37
28
1D

The entry is removed from the list and the ones below It are moved up one position. The
length of the list IS reduced by 1.

9-16

2)

Add an Entry to an Ordered List

Purpose:

Place the contents of memory location 0040 Into an ordered list if It IS not
already there. The length of the list IS in memory location 0041, and the list
itself beginS In memory location 0042, which consists of unsigned binary
numbers in increaSing order. Place the new entry in the correct position In
the list. adjust the elements below It down, and Increase the length of the
list by 1.

Sample Problems:

a.

Result:

b,

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)

6B
04
37
55
70
A1

(0041)
(0044)
(0045)
(0046)

05
6B
70
A1

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)

6B
04
37
55
6B
A1

Result: No change, since the entry IS already In the list.

3)

Add an Element to a Queue

Purpose:

Add the address In memory locations 0040 and 0041 (MSBs In 0041) to a
queue, The address of the first element of the queue is in memory locations
0042 and 0043 (MSBs In 0043). Each element in the queue contains either
the address of the next element in the queue or zero if there IS no next element; all addresses are 16 bits long with the most significant bits In the
second word of the element. The new element goes at the end {taill of the
queue; ItS address will be In the element that was at the end of the queue
and it will contain zero to Indicate that it is now the end of the queue,

Sample Problem:

(0040)
(0041)
(0042)
(0043)
(0046)
(0047)
Result;

(0046)
(0047)
(0040)
(004E)

6~}

new element to be added to queue

6~}

pointer to head of queue

gg} last element in queue

40} old last element points to
00
new last element

gg} new last element In queue

How would you add an element to the queue if memory locations 0044 and 0045 contained the address of the tail (last element) of the queue?

9-17

4)

16-Bit Sort

Purpose:

Sort an array of unsigned 16-blt binary numbers into descending order. The
length of the array IS m memory location 0040 and the array itself begins In
memory location 0041. Each 16-blt number is stored with the least significant bits In the first word.

Sample Problem:

Result:

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)
(0046)

03
01
19
60
3F
2A
B5

(0041)
(0042)
(0043)
(0044)
(0045)
(0046)

2A
B5
60
3F

Dl
19

The numbers are B52A. 3F60. and 1901

5)

Using a Jump Table With a Key

Purpose:

Use the contents of memory location 0040 as the key to a lump table starting In memory location 0041. Each entry in the Jump table contains an 8-blt
key value followed by a 16-bit address (MSBs m second word) to which the
program should transfer control if the key IS equal to that key value.

Sample Problem:

Result:

(0040)
(0041)
(0042)
(0043)
(0044)
(0045)
(0046)
(0047)
(0048)
(0049)
(PC)

38
32
4B
00
35
40
00
38
4F
00
004F. since that address corresponds
to key value 38.

Try wrltmg the program with and without the CPIR instruction. Can you think of a way
to simplify the version that uses the CPIR instruction? Hint: place all the corresponding
8-blt words into separate tables so that the pr~ram only has to add 1 to the table
pomter to move from one key value to the next.

9-18

REFERENCES
1.

Knuth describes other searching techniques In his book The Art of ComQuter Programming. Volume III: Sorting and Searching, Addison-Wesley, Reading, Mass"
1978. Knuth also has discussed searching and hashing In a more elementary way
In an article entitled"Algorithms" (see the April 1977 Issue of Scientific American).

2.

There are many sorting algorithms that vary wideIv in efficiency. Knuth describes
some in the book mentioned above (The Art of Comguter Programming. Volume
III: Sorting and Searching). Sorting and searching algorithms are also discussed In
K. A. Schember and J. R. Rumsey. "Minimal Storage Sorting and Searching Techniques for RAM Applications. a Tutorial". Comguter. June 1977. pp. 92-100.

3.

There are additional examples of the use of jump tables in L. A. Leventhal. "Cut
Your Processor's Computation Time". Electronic DeSign. August 16. 1977, pp.
82-89. and in Chapter 7 of J. B. Peatman, Microcomputer-Based Design.. McGrawHill. New York, 1977

4.

This method is discussed by T. Dollhoff In "Microprocessor Software: How to Optimize Timing and Memory Usage: Part Four: Techniques for the Zilog Z80".
Q!gltal Design.. February 1977. pp. 48-49.

9-19

Chapter 10
SUBROUTINES
None of the examples that we have shown so far IS typically a program all by Itself.
Most real programs perform a series of tasks. many of which may be the same or may
be common to several different programs. We need a way to formulate these tasks once
and make the formulations conveniently available both in different parts of the current
program and In other programs.

SUBROUTINE
The standard method IS to write subroutines that perform particuLIBRARY
lar tasks. The resulting sequences of instructions can be written
once. tested once. and then used repeatedly. They can form a
subroutine library that provides documented solutions to common problems.
SUBROUTINE
Most microprocessors have special instructions for transferring
control to subroutines and restoring control to the main proINSTRUCTIONS
gram. We often refer to the special Instruction that transfers
control to a subroutine as Call. Jump-to-Subroutine. Jump and Mark Place. or Jump
and Link. The special instruction that restores control to the main program is usually
called Return. On the
microprocessor. the Call Instruction (CALL) saves the old
value of the Program Counter In the RAM Stack before placing the starting address of
the subroutine into the Program Counter; the Return instruction (RET) gets the old
value from the Stack and puts it back In the Program Counter. The effect is to transfer
program control. first to the subroutine and then back to the main program. Clearly the
subroutine may itself transfer control to a subroutine. and so on.

zao

In order to be really useful. a subroutine must be general. A routine that can perform
only a specialized task. such as looking for a particular letter in an input string of fixed
length. will not be very useful. If. on the other hand. the subroutine can look for any letter in strings of any length. it will be far more helpful. We call the data or addresses that
the subroutine allows to vary "parameters" An Important part of writing subroutines is
deciding which variables should be parameters.

PASSING
One problem IS transferring the parameters to the subroutine: this
PARAMETERS
process is called passing parameters. The simplest method is for
the main program to place the parameters Into registers. Then the
subroutine can simply assume that the parameters are there. Of course. this technique
is limited by the number of registers that are available. The parameters may, however.
be addresses as well as data. For example, a sorting routine could begin with the startIng address of an array in Register Pair HL.
Other methods are necessary when there are more parameters. One possibility is to use
the Stack. The main program can place the parameters into the Stack and the
subroutine can retrieve them. The advantages of this method are that the Stack is essentially unlimited In size. and that data In the Stack IS not lost even if the Stack is used
again.

zao

The disadvantages are that few

instructions use the Stack. and the Call instruction
also stores the return address In the Stack. Another method is to use an area of memory
for parameters. The main program can place the address of the area Into Register Pair
HL or Into one of the Index registers and the subroutine can retrieve the data as needed.
However, this procedure is awkward if the parameters themselves are addresses.

10-1

Sometimes a subroutine must have special characteristics. A

RELOCATION
subroutine is relocatable if it can be placed anywhere in memory.
You can use such a subroutine easily. regardless of the placement of other programs or
the arrangement of the memory. A strictly relocatable program can use no absolute addresses: all addresses must be relative to the start of the program. A relocating loader
IS necessary to place the program in memory properly: the loader will start the program
after other programs and will add the starting address or relocation constant to all addresses in the program.

REENTRANT
A subroutine is reentrant if it can be interrupted and called by the
SUBROUTINE
interrupting program and still give the correct results for both the
interrupting and interrupted programs. Reentrancy is important for
standard subroutines In an interrupt-based system. Otherwise the Interrupt service
routines cannot use the standard subroutines without causing errors. Microprocessor
subroutines are easy to make reentrant. since the Call instruction uses the Stack and
that procedure is automatically reentrant. The only remaining requirement IS that the
subroutine use the registers and Stack rather than fixed memory locations for temporary storage. This IS a bit awkward. but usually can be done.if necessary.

A subroutine is recursive if it calls itself. Such a subroutine clearly must also be reentrant. However. recursive subroutines are uncommon in microprocessor applications.
Most programs consist of a main program and several subroutines. This IS advantageous because you can use proven routines and debug and test the other subroutines
separately. You must. however. be careful to use the subroutines properly and remember their exact effects on registers and memory locations.

SUBROUTINE DOCUMENTATION
Subroutine listings must provide enough information so that
users need not examine the subroutine's internal structure.
Among the necessary specifications are:

DOCUMENTING
SUBROUTINES

A description of the purpose of the subroutine.

A list of input and output parameters.
Registers and memory locations used.
A sample case.
If these guidelines are followed. the subroutine will be easy to use.

EXAMPLES
It IS important to note that the following examples all reserve an area of memory for the
RAM Stack. If the monitor invour microcomputer establishes such an area. you may use
it instead. If you wish to try establishing your own Stack area. remember to save and
restore the monitor's Stack Pointer in order to produce a proper return at the end of
your main program.
To save the monitor Stack POinter. use the Instruction LD (addr),SP. To restore the
monitor Stack Pointer. use the instruction LD SP.{addr). Both of these instructions require a two-bvte operation code (ED 78 for loading the Stack Pointer. ED 73 for storing
it) in addition to the two bytes of address.
We have used address 0080 (hex) as the starting point for the Stack. You may have to
consistently replace that address with one more suitable for your configuration. You
should consult your microcomputer's manual to determine the required changes.

10-2

Hex to ASCII
Purpose: Convert the contents of the Accumulator to an ASCII character. Place the
result in the Accumulator. Assume that the Accumulator contains a single
hexadecimal digit.
Sample Problems:
a.
Result:
b.
Result:

(A)

OC

(A)

43

(A)

06

(A)

36

'C'

'6'

Flowchart:

IA) = IAI 'ASCII AASCIl9-1

(AI = (AI + ASCII 0

Source Program:
The calling program starts the Stack at memory location 0080. gets the data from
memory location 0040. calls the conversion subroutine, and stores the result In memory
location 0041.
ORG
LD
LD
CALL
LD
HALT

o
SP,80H
A, (40H)
ASDEC
(41H),A

:START STACK AT LOCATION 0080

;GET DATA
:CONVERT DATA TO ASCII
:STORE RESULT

The subroutine converts a hexadeCimal digit to ASCII.

ORG

ASDEC:
ASCZ:

CP
JR
ADD
ADD

20H
10

:IS DATA A DECIMAL DIGIT?

C,ASCZ
A:A'-'9'-1
A:O'

;NO, ADD OFFSET FOR LETTERS

;CONVERT DATA TO ASCII

Rn

10-3

Subroutine Documentation:
, SUBROUTINE ASDEC
PURPOSE: ASDEC CONVERTS A HEXADECIMAL
DIGIT IN THE ACCUMULATOR TO AN
ASCII DIGIT IN THE ACCUMULATOR
. INITIAL CONDITIONS: HEX DIGIT IN A
: FINAL CONDITIONS: ASCII CHARACTER IN A
. REGISTERS USED: A
SAMPLE CASE
INITIAL CONDITIONS: 6 IN ACCUMULATOR
FINAL CONDITIONS, ASCII 6 (HEX 36)
IN ACCUMULATOR
Object Program:
Memory Address
(Hex)
1)

Instruction
{Mnemonic}

Calling program
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
0008
OOOC

2)

Memory Contents
(Hex)

31
80
00
3A
40
00
CD
20

LD

SP,80H

LD

A.(40H)

CALL

ASDEC

LD

(41Hl,A

00
32
41
00
76

HALT

Subroutine
0020
0021
0022
0023
0024
0025
0026
0027
0028

ASDEC.

FE
OA
38
02
C6
07
C6
30
C9

ASCZ:

CP

10

JR

C,ASCZ

ADD

A:A'-'9'-1

ADD

A:O'

RET

The Instruction LD SP.80H starts the Stack at memory location 0080. Remember that
the Stack grows downward {to lower addresses!' We usually place the Stack at the high

10-4

end of RAM (i.e.. the highest address) so that It will not Interfere with other temporary
storage.
The Call instruction places the subroutine starting address (0020 hex) Into the Program
Counter and saves the old Program Counter (0009 hex) In the Stack. The procedure IS:
STEP 1 -

Decrement Stack POinter. save MSBs of old Program Counter in Stack.

STEP 2 -

Decrement Stack Pointer. save LSBs of old Program Counter in Stack.

Note that the Z80 Stack POinter always contains the address of the last occupied Stack
location.
The result In this case IS:
(OOlF)
(OOlE)
(SP)

00
09
OOlE

The value that IS saved IS the value of the Program Counter after the processor has
fetched the entire Call instruction from memory. Note that the address ends up stored
lust like other Z80 addresses. with the least significant bits In the lower address.
The Return instruction loads the Program Counter with the contents of the bottom two
memory locations In the Stack. The procedure IS:
STEP 1 -

Load eight bits from Stack Into LSBs of Program Counter. Increment Stack
POinter.

STEP 2 -

Load eight bits from Stack into MSBs of Program Counter. Increment Stack
POinter.

The result In this case IS:

(PC)
(SP)

(OOlF) and (OOlE)

0009
0080

This subroutine has a single Input parameter and produces a single result. The Accumulator IS the obvious place to put both.
The calling program Involves three steps: plaCing the data Into the Accumulator. callIng the subroutine. and stOring the result. The overall initialization must also place the
Stack In the appropriate area of memory.
The subroutine is reentrant. since It uses no data memory; It IS relocatable. since the
address ASCZ is relative.
Note that the CALL instruction results In the execution of four or five instructions taking
36 or 38 clock cycles. A subroutine call can take a long lime even though It appears to
be a Single instruction In the program.
If you plan to use the Stack for parameters. remember that CALL places the return address at the top of the Stack. You can increment the Stack POinter tWice (INC SP) to get
past the return address. but you must also remember to adjust the Stack POinter properly before returning. You can also move the Stack POinter to Registers Hand L With
the sequence:
LD
ADD

HL.O
HL.SP

;MOVE STACK POINTER TO ADDRESS REGISTER

Now yOU can use Implied memory addreSSing With Hand L to ac-cess data in the Stack.
Another alternative IS to move the Stack POinter to an Index register (say IX) with the sequence:
LD
ADD

IX.O
IX.SP

;MOVE STACK POINTER TO INDEX REGISTER

10-5

This alternative has the advantage that you can now access data and addresses in the
Stack with Indexed offsets. Furthermore, Register Pair HL is Immediately available for
use in the subroutine. Note that you can use the instructions LD SP,HL or LD SP,IX to
return an adiusted value to the Stack POinter.

Length of a String of Characters

Purpose: Determine the length of a string of ASCII characters. The starting address of
the string is In Register Pair HL The end of the string IS marked by a carnage
return character (CA. hex DOl. Place the length of the string (excluding the
carnage return) into the Accumulator.
Sample Problems:
(HU
(0043)

a.
Result:

0043
00

(A)

00

(HU
(0043)
(0044)
(0045)
(0046)
(0047)
(0048)
(0049)

b.

Result:

0043
52 'R'
41 'A'
54 'T'
48 'H'
45 'E'
52 'R'
00 CR

(A)

06

Flowchart:

Pointer

Count

HL

Ves

Count

Count + 1

PQlnter

Pomter + ,

10-6

Source Program:
The calling program starts the Stack at memory location 0080. gets the starting address
of the string from memory locations 0040 and 0041. calls the string length subroutine.
and stores the result in memory location 0042.
LD
LD
CALL
LD
HALT

SP.80H
HL.(40H)
STLEN
(42Hl.A

:START STACK AT LOCATION 0080

:GET STARTING ADDRESS OF STRING
:DETERMINE STRING LENGTH
:STORE STRING LENGTH

The subroutine determines the length of a string of ASCII characters and places the
length Into the Accumulator.
STLEN:
CHKCR:

20H
B.O
AODH
(HL)
Z.DONE

ORG
LD
LD
CP

JR

INC
INC

HL
CHKCR
A.B

JR
DONE:

ZERO
:STRING LENGTH
:GET ASCII CARRIAGE RETURN
:IS CHARACTER A CARRIAGE RETURN?
:YES. END OF STRING
:NO. ADD 1 TO STRING LENGTH

LD
RET

Subroutine Documentation:
:SUBROUTINE STLEN
:PURPOSE. STLEN DETERMINES THE LENGTH OF A
STRING (NUMBER OF CHARACTERS BEFORE
A CARRIAGE RETURN)
:INITIAL CONDITIONS. STARTING ADDRESS OF
STRING IN REGISTER PAIR HL
:REGISTERS USED: A.B.H.L
;SAMPLE CASE:
ST ARTING CONDITIONS: (HL) = 0043
(0043)

= 35.

(0044)

FINAL CONDITIONS

= 46.

(0045)

= 00

(A) = 02

10-7

Object Program:
Memorv Address
(Hex)
1)

Instruction
(Mnemonic)

Calling program
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC

2)

Memorv Contents
(Hex)

31
80
00
2A
40
00
CD
20
00
32
42
00
76

LD

SP.80H

LD

HL.(40H)

CALL

STLEN

LD

(42Hl.A

HALT

Subroutine
0020
0021
0022
0023
0024
0025
0026
0027
0028
0029
002A
002B
002C

06
00
3E
OD
BE
28
04
04
23
18
F9
78
C9

STLEN:

CHKCR:

DONE:

LD

B.O

LD

AODH

CP
JR

(HU
Z.DONE

INC
INC
JR

B
HL
CHKCR

LD
RET

A.B

The calling program Involves four steps: initializing the Stack POinter. placing the startIng address of the strrng Into Register Pall HL. calling the subroutine. and storrng the
result.
The subroutine is reentrant. since it does not change the contents of any memory locations. It is relocatable. since all the Jump instructions use relative addresses.
The subroutine changes Register B and the address in Register Pair HL as well as the
Accumulator. The programmer must be aware that data previously stored in Register B
and the address previously loaded into HL will be lost: the subroutine documentatIOn
must describe what registers are used.
An alternative to destroYing register contents In the subroutine IS to save them In the
Stack and then restore them before returning. This approach makes the calling routine
Simpler. but costs extra time and memory (in the program and In the Stack).
This subroutine has a single Input parameter. which is an address. The best way to pass
thiS parameter IS through a register pall and. since the HL pall is certainly the most flexible as far as addressing options are concerned. it is the obvIous choice.
The subroutine contains an unconditional Jump instruction. JR CHKCR. By alterrng the
initial conditions prior to entering the subroutlne's loop. can yOU eliminate this iump?
If the terminating character were not always an ASCII carriage return. we could make
that character into another parameter. Now the calling program would have to place

10-8

the terminating character into the Accumulator and the starting address of the string
Into Register Pair HL before calling the subroutine.
One way to pass parameters that do not depend on vanable data is to place the values
In program memory immediately after the Call Instruction. You can use the old Program
Counter (saved at the top of the Stack) to access the data. but you must adjust itS value
properly before returning control to the main program. For example. we could pass the
value of the terminating character this way. The main program and subroutine would
be:
Calling program:
ORG
LD
LD
CALL
DEFB
LD
HALT

0
SP.80H
(HU.40H
STLEN
(42H).A

:START STACK AT LOCATION 0080

:GET STARTING ADDRESS OF STRING
:DETERMINE STRING LENGTH
.TERMINATOR = ASCII PERIOD
:STORE STRING LENGTH

Subroutine:
STLEN:

CHKCR:

DONE:

ORG
POP
LD
INC
PUSH
LD
CP
JR
INC
INC
JR
LD
RET

20H
DE
A.(DE)
DE
DE
B.O
(HU
Z.DONE
B
HL
CHKCR
A.B

:GET START OF PARAMETER LIST

:GET TERMINATING CHARACTER
:ADJUST RETURN ADDRESS
:STRING LENGTH = ZERO
:IS CHARACTER TERMINATOR?
:YES. END OF STRING
:NO. ADD 1 TO STRING LENGTH

This subroutine is longer and uses Register Pair DE. but the calling program need not
load the terminating character Into a register. The INC DE instruction is necessary to
force a return to the next instruction. rather than to the parameter list. 1
PUSH and POP transfer the contents of register pairs or index registers to and from the
RAM Stack. The eight least significant bits are removed first and stored last to retain
consistency With the Z80's upside-down method of storing 16-bit addresses. Remember that the RAM Stack grows downward (to lower addresses!.

Add Even Parity to ASCII Characters

Purpose: Add even parity to a string of 7-bit ASCII characters. The length of the stnng
is in the Accumulator and the starting address of the string is in Register Pair
HL Place even panty in the most significant bit of each character. i.e.. set the
most significant bit to 1 if that makes the total number of 1 bits In the wore'

even.

10-9

Sample Problem:
(A)

(0041)
(0042)
(0043)
(0044)
(0045)
(0046)

06
0041
31
32
33
34
35
36

(0041)
(0042)
(0043)
(0044)
(0045)
(0046)

81
82
33
84
35
36

(HLl

Result:

Flowchart:

POinter

fHLl

Count

lA)

(Pointer)

(POinter)

OR lOOOOOOOB
(Set Panty Bid

POinter

=Pomter

Count

= Count

10-10

+ 1
- 1

Source Program:
The calling program starts the Stack at memory location 0080, sets the starting address
of the string to 0041, gets the string length from memory location 0030, and calls the
even parity subroutine.
ORG
LD
lD
LD
CALL
HALT

o
SP.80H
HL.41H
A,(30H)
EPAR

:START STACK AT LOCATION 0080

:GET STARTING ADDRESS OF STRING
:GET STRING LENGTH

The subroutine adds even parity to a string of ASCII characters.

EPAR:
SETPR:

CHCNT

ORG
LD
lD
LD
OR
JP
LD
INC
DJNZ
HALT

20H
B,A
C.l0000000B
A,(HL)
C
PO.CHCNT
(HL),A
HL
SETPR

:GET PARITY BIT OF 1

:GET A CHARACTER
:SET PARITY BIT TO 1
:IS PARITY NOW EVEN?
:YES. SAVE CHARACTER WITH EVEN PARITY

Subroutine Documentation:
:SUBROUTINE EPAR
:PURPOSE: EPAR ADDS EVEN PARITY
TO A STRING OF 7-BIT ASCII
. CHARACTERS
:INITIAL CONDITIONS: STARTING ADDRESS
OF STRING IN HL. LENGTH OF STRING
. IN A
:FINAL CONDITIONS: EVEN PARITY IN
MSB OF EACH CHARACTER
:REGISTERS USED: A,B,C.H,L
:SAMPLE CASE:
INITIAL CONDITIONS: (HL) = 0041
(A) = 2. (0041) = 32, (0042) = 33
FINAL CONDITIONS: (0041) = B2, (0042) = 33
This subroutine has two parameters, an address and a number. Register Pair HL IS used
to pass the address and the Accumulator to pass the number. No explicit results are
returned, since the subroutine affects only the MSB of each character In the string.

10-11

The calling program must place the starting address of the string mto Register Pair HL
and the length of the string mto the Accumulator before transferring control to the
subroutine.
The subroutine changes the values In Registers A. H. and L and uses Registers Band C
for temporary storage. It IS reentrant. since it does not use any fixed memory locations
for temporary storage.
Object Program:
Memory Address
(Hex)
1)

Instruction
(Mnemonic)

Calling program
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC

2)

Memory Contents
(Hex)

31
80
00
21
41
00
3A
30
00
CD
20
00
76

LD

SP.80H

LD

HL,41 H

LD

A.(30H)

CALL

EPAR

HALT

Subroutine
0020
0021
0022
0023
0024
0025
0026
0027
0028
0029
Q02A
002B
002C

47
OE
80
7E
Bl
E2
29
00

EPAR:

LD
LD

B.A
C.l0000000B

SETPR:

LD
OR
JP

A,(HL)
C
PO.CHCNT

LD
INC
DJNZ

(HL),A
HL
SETPR

77
23
10
F7
C9

CHCNT

RET

Pattern Match
Purpose: Compare two strings of ASCII characters to see if they are the same. The
length of the strings is in the Accumulator. The starting address of one string
IS in Register Pair HL: the starting address of the other IS in Register Pair DE.
If the two strings match, clear the Accumulator: otherwise. set theAccumulator to FF (hex).

10-12

Sample Problems:
(A)

a.

Result:

b.

Result:

(DE)
(HU

03
50
60

(0050)
(0051)
(0052)

43
41
54

'C'
'A'

(0060)
(0061)
(0062)

43
41
54

'C'
'A'

(A)

0, since the strings are the same.

(A)

(DE)
(HU

03
50
60

(0050)
(0051)
(0052)

52
41
54

(0060)
(0061)
(0062)

43
41
54

(A)

'R'

'A'

T
'C'
'A'

FF (hex), since the first characters differ.

Flowchart:

Pointer 1
Pointer 2
Count

fDEI
fHLI
fA)

No

Pomter 1. ;;
Pointer 1 + 1
Pomter 2 =
Pointer 2 + 1
Count = Count - 1

fA)=O

10-13

fA) = FF fhe.)

Source Program:
The calling program starts the Stack at memory location 0080. sets the starting addresses of the stnngs to 0050 and 0060. respectively. gets the string length from
memory location 0040. calls the pattern match subroutine. and places the result into
memory location 0041.
ORG
LD
LD
LD
LD
CALL
LD
HALT

o
SP.80H
DE.60H
HL.50H
A.(40H)
PMTCH
(41Hl.A

:START STACKAT LOCATION 0080

:GET STARTING ADDRESS OF STRING 1
:GET STARTING ADDRESS OF STRING 2
:GET STRING LENGTH
:CHECK FOR MATCH
:SAVE MATCH INDICATOR

The subroutine determines if the two strings are the same.

PMTCH:
CHCAR:

DONE.

ORG
LD
LD
LD
CP
JR
INC
INC
DJNZ
LD
LD
RET

20H
8.A
C,OFFH
A.(DE)
(HU
NZ,DONE
DE
HL
CHCAR
C,O

:COUNT = STRING LENGTH

:MARK = FF (HEX) FOR NO MATCH
:GET CHARACTER FROM STRING 1
:IS THERE A MATCH WITH STRING 27
:NO, DONE-STRINGS D.O NOT MATCH

:MARK = ZERO. STRINGS MATCH

AC

Subroutine Documentation:
:SUBROUTINE PMTCH
:PURPOSE: PMTCH DETERMINES IF TWO
.

STRINGS ARE EQUIVALENT

:INITIAL CONDITIONS: STARTING ADDRESSES

OF STRINGS IN DE AND HL,
LENGTH OF STRINGS IN ACCUMULATOR
:FINAL CONDITIONS: 0 IN A IF
STRINGS MATCH. FF IN A OTHERWISE
:REGISTERS USED: A.B,D,E.H,L
:SAMPLE CASE:
STARTING CONDITIONS: (HU = 0050.
(DE) = 0060. (A) = 2
(0050) = 36, (0051) = 39
(0060) = 36, (0061) = 39
FINAL CONDITIONS: (A) =0 SINCE THE STRINGS MATCH

10-14

Object Program:

Memory Address
(Hex)
1)

Instruction
(Mnemonic)

Calling program
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012

2)

Memory Contents
(Hex)

31
80
00
11
60
00
21
50
00
3A
40
00
CD
20
00
32
41
00
76

LD

SP,80H

LD

DE,60H

LD

HL,50H

LD

A.(40H)

CALL

PMTCH

LD

(41H!.A

HALT

Subroutine
0020
0021
0022
0023
0024
0025
0026
0027
0028
0029
002A
002B
002C
0020
002E

47
OE
FF
1A
BE
20
06
13
23
10
F8
OE
00
79
C9

PMTCH:

LD
LD

B,A
C,OFFH

CHCAR:

LD
CP
JR

A,(DE)
(HU
NZ,DONE

INC
INC
DJNZ

DE
HL
CHCAR

LD

C,O

LD
RET

A.C

DONE:

This subroutine, like the preceding ones, changes all of the flags. You should generally
assume that a subroutine call changes the flags unless it IS specifically stated otherwise. If the main program needs the old flag values (for later checking!. it must save
them In the Stack prior to calling the subroutine. This is accomplished with the PUSH
AF instruction.
The subroutine is reentrant and changes all the main registers except C.
This subroutine has three parameters - the two starting addresses and the length of
the strings. These parameters use five general-purpose registers.

10-15

Multiple-Precision Addition
Purpose: Add two multiple-byte binary numbers. The length of the numbers in bytes
is In the Accumulator. The starting addresses of the numbers are in Register
Pairs DE and HL. The starting address of the result IS in Index Register IX. All
the numbers begin with the least Significant bits.
Sample Problem:
(A)
(DE)

(IX)

04
51
61
71

(0051)
(0052)
(0053)
(0054)
(0061)
(0062)
(0063)
(0064)

C3
A7
58
2F
8S
35
DF
14

Result:

(0071)
(0072)
(0073)
(0074)

(HL)

I.e.

= 78
= DD
=3A
=44

2F58A7C3
14DF358S
443ADD78

Flowchart:

Count

POinter 1
POinter 2
POinter 3

(DEI
(HLI
(IX)'

-ea

(POinter 3)

(Pointer 1)
+ (Pointer 2)

+ earrv
---"l!""--'"
Pointer 1::: Pointer 1+1
POinter 2;= POinter2+1

Pointer 3== POlnter3+1

Count = Count - l

10-16

(lhis step also produces new carry)

Source Program:
The calling program starts the Stack at memory location 0080, sets the starting addresses of the vanous numbers to 0050,0060, and 0070, respectively, gets the length
of the numbers from memory location 0040, and calls the multiple-precision addition
subroutine.
ORG
LD
LD
LD
LD
LD
CALL
HALT

o
;START STACK AT LOCATION 0080
:GET STARTING ADDRESS OF FIRST NUMBER
:GET STARTING ADDRESS OF SECOND NUMBER
:GET STARTING ADDRESS OF RESULT
:GET LENGTH OF NUMBERS IN BYTES
:MULTIPLE-PRECISION ADDITION

SP.80H
HL,50H
DE.60H
IX,70H
A,(40H)
MPADD

The subroutine performs multiple-precision binary addition.

ORG
MPADD; LD
AND
ADDW: LD
ADC
LD
INC
INC
INC
DJNZ
RET

20H
B,A

:COUNT = LENGTH OF NUMBERS IN BYTES

:CLEAR CARRY TO START
:GET WORD FROM FIRST NUMBER
:ADD WORD FROM SECOND NUMBER
:STORE ONE WORD OF RESULT

A
A,(DE:
A,(HL)
(lXl.A
DE
HL
IX
AD OW

Subroutine Documentation:
:SUBROUTINE MPADD
:PURPOSE. MPADD ADDS TWO
.

MULTIPLE-BYTE BINARY NUMBERS

:INITIAL CONDITIONS; STARTING ADDRESSES

OF NUMBERS IN D AND E, HAND L.
STARTING ADDRESS OF RESULT IN IX,
LENGTH OF NUMBERS IN A
;REGISTERS USED; A,B,D,E,H,L.IX
;SAMPLE CASE:
STARTING CONDITIONS. (HL) = 0050,
(DE) = 0060, (IX) = 0070, (A) = 2,
(0050)

= C3.

(0051)

= A7,

(0060)

= B8,

(0061)

FINAL CONDITIONS: (0070) = 7B, (0071) = DO

10-17

= 35

Object Program"

Memory Address
(Hex)
1)

Instruction
(Mnemonic)

Calling program
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012
0013

2)

Memory Contents
(Hex)
31
80
00
21
50
00
11
60
00
DD
21
70
00
3A
40
00
CD
20
00
76

LD

SP.80H

LD

HL,50H

LD

DE.60H

LD

IX,70H

LD

A,(40H)

CALL

MPADD

HALT

Subroutine
0020
0021
0022
0023
0024
0025
0026
0027
0028
0029
002A
002B
002C
002D

47
A7
lA
8E
DO
77
00
13
23
DD
23
10
F5
C9

LD
AND
LD
ADC
LD

B,A
A
A,(DE)
A,(HL)
(IXl.A

INC
INC
INC

DE
HL
IX

DJNZ

ADDW

RET

We use Index Register IX to hold the result address. Try changing the program to use
Register Pair BC for this purpose. What happens to the counter?
We could also place the result address at the top of the Stack. The instruction EX
(SPl.HL exchanges the top of the Stack and Register Pair HL. Change the program so
that it uses thiS instruction, but remember to increment all three pOinters after each
Iteration.
This subroutine has four parameters - three addresses and the length of the numbers.
Six 8-blt registers and the 16-bit Index Register IX are used for passing parameters.

10-18

PROBLEMS
Note that you are to write both a calling program for the sample problem and a properly
documented subroutine.

1)

ASCII to Hex

Purpose: Convert the contents of the Accumulator from the ASCII representation of a
hexadecimal digit to the 4-blt binary representation of the digit. Place the
result Into the Accumulator.
Sample Problems:

a.
Result:
b.
Result:

2)

(A)

43

(A)

OC

(A)

36

(A)

06

'C'
'6'

Length of an ASCII Message

Purpose: Determine the length of an ASCII-coded message. The starting address of

the string of characters in which the message IS located is In Register Pair
HL The message Itself starts with an ASCII STX character (hex 02) and ends
with ASCII ETX (hex 03). Place the length of the message (the number of
characters between the STX and the ETX) into the Accumulator.
Sample Problem:
(Hl)
(0041)
(0042)
(0043)
(0044)
(0045)

Result:

3)

(A)

0041
49
02 STX
47 'G'
4F '0'
03 ETX
02

Check Even Parity in ASCII Characters

Purpose: Check the even parity of a string of ASCII characters. The length of the string
is In the Accumulator and the starting address of the string is in Register Pair
HL If the panty of all the characters in the string IS correct clear the Accumulator: otherwise, set the Accumulator to FF hex (all ones).
Sample Problems:

a.

(A)
(Hl)
(0042)
(0043)
(0044)

Result:

b.

82
33

(A)

00, since all the characters have even parity

(A)

03
0042
81

(HU
(0042)
(0043)
(0044)
Result:

03
0042
81

(A)

86
33
FF, since the character In memory location 0043
does not have even parity

10-19

4)

String Comparison

Purpose: Compare two strings of ASCII characters to see which IS larger (i.e., which
would follow the other in 'alphabetlcal' ordering).

The length of the strings IS In the Accumulator; the starting address of

string 1 is In Register Pair HL and the starting address of string 2 is In
Register Pair DE. If string 1 IS larger than or equal to string 2. clear the Accumulator; otherwise, set the Accumulator to FF hex (all ones).
Sample Problems:

a,

b.

c.

5)

(A)
(DE)
(HL)
(0050)
(0051)
(0052)

03
0060
0050
43 'C'
41 'A'
54 T

(0062)
(0063)
(0064)

42
41
54

Result

(A) = 00, since CAT IS 'larger than BAT

(A)
(DE)
(HL)
(0050)
(0051)
(0052)

03
0060
0050
44 'D'
4F '0'
47 'G'

(0060l
(0061)
(0062)

44
4F
47

'B'
'A'

'D'
'0'
'G'

Result

(A) = 00, since the two strings are equal

(A)
(DE)
(HU
(0050)
(0051)
(0052)

03
0060
0050
43 'C'
41 'A'
54 T

(0060)
(0061)
(0062)

43
55
54

Result

(A) = FF (hex!. since CUT is 'larger' than CAT

'C'
'U'

Decimal Subtraction

Purpose: Subtract one multiple-digit decimal (BCD) number from another. The length
of the numbers (in bytes) IS In the Accumulator and the starting addresses of
the numbers are In Register Pairs DE and HL. Subtract the number with the
starting address In HL from the one with the starting address In DE. The
starting address of the result is in Index Register IX. All the numbers begin
with the least significant digits. The sign of the result IS returned In the Accumulator - zero if the result IS positive. FF (hex) if It is negative.

10-20

Sample Problem:
(A)

04
0050
0060
0070

(DE)
(HU
(IX)

(0050)
(0051)
(0052)
(0053)
(0060)
(0061)
(0062)
(0063)
Result:

85
19
70
36
59
34
66
12
00 (positive)
26
85
03
24

(A)

(0070)
(0071)
(0072)
(0073)

36701985
12663459

I.e..

24038526

10-21

REfERENCES
1.

Other examples of this technique (for the 8080 microprocessor) are In S. Mazor and
C. Pitchford, "Develop Cooperative Microprocessor Subroutines," Electronic
Design, June 7,1978, pp. 116-118.

10-22

Chapter 11
INPUT/OUTPUT
There are two problems In the design of input/output sections: one IS how to interface
peripherals to the computer and transfer data. status. and control signals: the other IS
how to address I/O devtces so that the CPU can select a particular one for a data
transfer. Clearly. the first problem is both more complex and more interesting. We will
therefore discuss the Interfacing of peripherals here and leave addressing to a more
hardware-oriented book.
In theory. the transfer of data to or from an I/O device is similar to
the transfer of data to or from memory. In fact. we can consider the
memory as Just another I/O device. The memory is. however.
special for the following reasons:
1)
2)
3)
4)
5)

It operates at almost the same speed as the processor.

It uses the same type of signals as the CPU. The only Circuits usually needed to interface the memory to the CPU are drivers. receivers. and level translators.
It requires no special formats or any control signals besides a ReadlWrite pulse.
It automatically latches data sent to it
Its word length is the same as the computers.

Most I/O devices do not have such convenient features. They may operate at speeds
much slower than the processor: for example. a teletypewriter can transfer onlY 10
characters per second, while a slow processor can transfer 10.000 characters per second. The range of speeds is also very wide - sensors may provide one reading per
minute. while video displays or floppy disks may transfer 250.000 bits per second.
Furthermore. I/O devices may require continuous signals (motors or thermometers), currents rather than voltages (teletypewriters), or voltages at far different levels than the
signals used by the processor (gas-discharge displays). I/O devices may also require
special formats. protocols. or control signals. Their word lengths may be much shorter
or much longer than the word length of the computer. These variations make the
deSign of I/O sections difficult and mean that each peripheral presents its own special
interfacing problem.
We may. however, provide a general description of devices and interfacing methods. We may roughly separate devices into three
categories, based on their data rates:
1)

2)

3)

Slow devices that change state no more than once per second. Changing their
states typically requires milliseconds or longer. Such devices Include lighted displays. sWitches. relays. and many mechanical sensors and actuators.
Medium-speed devices that transfer data at rates of 1 to 10.000 bits per second.
Such devices Include keyboards, printers, card readers. paper tape readers and
punches. cassettes, ordinary communications lines, and many analog data acquiSItion systems.
High-speed deVices that transfer data at rates of over 10.000 bits per second. Such
devices include magnetic tapes. magnetic disks. high-speed line printers, hlghspeed communications lines, and video displays.

11-1

The Interfacing of slow devices is simple. Few control signals

INTERFACING
are necessary unless the devices are multiplexed. i.e.. several
SLOW DEVICES
are handled from one port. as shown in Figures 11-1 to 11-4.
Input data from slow devices need not be latched. since it remains stable for a long time
interval. Output data must of course. be latched. The onlv problems with input are
transitions that occur while the computer is reading the data. One-shots. cross-coupled
latches. or software delay routines can smooth the transitions.
A single port can handle several slow devices. Figure 11-1 shows a demultiplexer that
automaticallv directs the next output data to the next device bv counting output operations. Figure 11-2 shows a control port that provides select inputs to a demultiplexer.
The data outputs here can come in any order. but an additional output Instruction is
necessary to change the state of the control port. Output demultiplexers are commonly
used to drive several displays from the same output port. Figures 11-3 and 11-4 show
the same alternatives for an input multiplexer.
Note the differences between input and output with slow devices:
1)

2)
3)

Input data need not be latched. since the input device holds the data for an enormous length of time by computer standards. Output data must be latched. since
the output device will not respond to data that is present for only a few CPU clock
cycles.
Input transitions cause problems because of their duration; brief output transitions
cause no problems because the output devices (or the observers) react slowly.
The major constraints on input are reaction time and responsiveness. the major
constraints on output are response time and observability.

Medium-speed devices must be synchronized in some way to

INTERFACING
the processor clock. The CPU cannot simplv treat these devices
MEDIUMSPEED
as if they held their data forever or could receive data at any
DEVICES
time. Instead. the CPU must be able to determine when a
device has new Input data or is ready to receive output data. It must also have a way of
telling a device that new output data IS available or that the previous input data has
been accepted. Note that the peripheral may be or contain another processor.

The standard unclocked procedure is the handshake. Here the

HANDSHAKE
sender Indicates the availability of data to the receiver and
transfers the data; the receiver completes the handshake by acknOWledging the reception of the data. The receiver may control the situation by initially requesting the data or
by indicating ItS readiness to accept data; the sender then sends the data and completes the handshake by Indicating that data is available. In either case. the sender
knows that the transfer has been completed successfully and the receiver knows when
new data is available.

11-2

Data Outputs 0
)

Data Bus

Data

Output

Inputs

Port

Data Outputs 1

Strobe

)
y

Port SelectIon Logic

Demultiplexer
Data Outputs 2

>

Clock
y

Select

Counter

Inputs

Data Outputs 3

The Counter controls where the Demultiplexer sends the data.

Figure 11-1. An Output Demultiplexer Controlled by a Counter

..

Data Outputs 0

Data

Data
Port

.Inputs

Data Outputs 1

>

Oata Bus
Demultiplexer

Data Outputs 2

Control
Port

Select
Inputs

..

Data Outputs 3
)

The- CPU sends control information to the Control Port; that port detennjnes
where the Demultiplexer sends the data.

Figure 11-2. An Output Demultiplexer Controlled by a Port

11-3

Oata Inputs 0
Data Bus

Oata
Outputs

Input

Port

Port SeIec:tion Logic

Data Inputs 1
~

Multiplaxar
A

Data Inputs 2

Cock
~

Salact

Countar

Data Inputs 3

Inputs

The Countar controls which input the Multiplexer lJllleS to the Input Port.

Figure 11-3. An Input MUltiplexer Controlled by a Counter

Dat8 Inputs 0

Input Data
""~

Bus

Data

Data

Port

Outputs
A

Data Ihputs 1

Data Inputs 2

Multiplexer

.l

Output Data Bus

i

Selsct

Control
Port

Inputs

.. Oata Inputs 3
;/
~

The control information which the CPU sends to the Control Port (with an output oparationi
determines which input the Muhiplexar routes to the Oata Port.

Figure 11-4. An Input Multiplexer Controlled by a Port

11-4

Figures 11-5 and 11-6 show typical input and output operations using the handshake
method. The procedure whereby the CPU checks the readiness of the peripheral before
transferring data is called "polling" Clearly. polling can occupy a large amount of processor time if there are many I/O devices. There are several ways of providing the
handshake signals. Among these are:
Separate dedicated I/O lines. The processor may handle these as additionalI/O ports
or through special lines or interrupts. The
processor does not have serial I/O lines.
but the
ParaliellnputiOutput device' (or PIO) does.
Special patterns on the I/O lines. These may be single start and stop bits or entire
characters or groups of characters. The patterns must be easy to distinguish from
background nOise or inactive states.

zao

zao

We often call a separate I/O line that indicates the availability of

data or the occurrence of a transfer a "strobe" A strobe may. for
example. clock data into a latch or fetch data from a buffer.

ISTROBE I

Many peripherals transfer data at regular intervals; i.e.. synchronously. Here the only
problem is starting the process by lining up to the first Input or marking the first output.
In some cases. the peripheral provides a clock input from which the processor can obtain timing information.
Transmission errors are a problem with medium-speed devices.
Several methods can lessen the likelihood of such errors; they
include:

REDUCING
TRANSMISSION
ERRORS

Sampling input data at the center of the transmission interval

in order to avoid edge effects; that IS. keep away from the edges where the data is
changing.
Sampling each input several times and using majority logic such as best three out of
five. 1
Generating and checking parity; an extra bit is used that makes the number of 1 bits
in the correct data even or odd.
Using other error detecting and correcting codes such as checksums. LRC
(longitudinal redundancy check). and CRC (cyclic redundancy check).2
~=~~~
High-speed devices that transfer more than 10.000 bits per
INTERFACING
second require special methods. The usual technique is to conHIGHSPEED
struct a special-purpose controller that transfers data directly
DEVICES
between the memory and the I/O device. This process is called
DIRECT
direct memorv access (DMA). The DMA controller must force
MEMORY
the CPU off the busses. provide addresses and control signals
ACCESS
to the memorv. and transfer the data. Such a controller will be
fairly complex. typically consisting of 50 to 100 chips.
although LSI devices are now available. 3 The CPU must initially load the Address and
Data Counters in the controller so that the controller will know where to start and how
much to transfer.

11-5

Input
Acknowledge

Data

Data Bus

CPU
~

Data Readv

a)

Peripheral provides data and Data Ready signal to computer I/O section.

Input

Acknowledge

Data

Data Bus

Penpheral

Data Ready

bi

CPU reads Data Readv signal from I/O section (this may be a hardware interrtJpt connection).

Input

Acknowledge

Data

Data Bus

Peripheral

Data Ready

ci

CPU reads data from I/O section,

Input

Acknowledge

Data Bus

Data

Data Ready

d)

CPU sends Input Acknowledge signal to I/O sectIon, which then provides Input Acknowledge signal
to Peripheral (this may be a hardware connection).

Figure 11-5. An Input Handshake

11-6

......._ ......_ _-,Output Ready

Data Bus

Data

CPU

al

Penpheral provides Penpheral Readv sIgnal to computer I/O sectIon.

Output Ready
Data Bus

Data
Penpheral
Peripheral Ready

bi

CPU reads Penpheral Ready sIgnal from 110 section (this may be a hardware Interrupt connection),

Output Ready
Data Bus

Data

"
Peripheral Ready

ci

CPU sends data to Penpheral.

Output Ready
Data Bus

Data

.
Peripheral Ready

d)

CPU sends Output Ready signal to Penpheral (this may be a hardware connection).

Figure 11-6. An Output Handshake

11-7

TIMING INTERVALS ([)ElAYS)

USES OF
One problem that we will face throughout the discussion of InTIMING
put/output is the generation of timing intervals with specific
INTERVALS
lengths. Such intervals are necessary to debounce mechanical
switches (to smooth their Irregular transitions). to provide pulses
with specified lengths and frequencies for displays. and to provide timing for devices
that transfer data regularly (for example. a teletypewriter that sends or receives one bit
everv 9.1 msl.
METHODS
FOR
In hardware with one-shots or monostable multivibrators.
PRODUCING
These devices produce a single pulse of fixed duration in
TIMING
response to a pu Ise input.
INTERVALS
In a combination of hardware and software with a flexible proCounter-Timer Circuit (or
grammable timer such as the
CTC) for Z80 based microcomputers. as described in An Introduction to Microcomputers: Volume 2 - Some Real Microprocessors. The CTC can provide timing intervals of various lengths with a variety of starting and ending conditions.
In software with delay routines. These routines use the processor as a counter. This
is possible since the processor has a stable clock reference. but it clearly under-utilizes the processor. However. delay routines require no additional hardware and
often use processor time that would otherwise be wasted.

We can produce timing Intervals in several wavs:

1)

2)

3)

zao

The choice among these three methods depends on your application. The software method is Inexpensive but may overburden the
processor. The programmable timers are relatively expensive. but
are easy to Interface and may be able to handle many complex
timing tasks.

CHOOSING

A TIMING
METHOD

DELAY ROUTINES
A simple delay routine works as follows:
Step 1)
Step 2)
Step 3)

Load a register with a specified value.

Decrement the register.
If the result of Step 2 is not zero. repeat Step 2.

BASIC
SOFTWARE
DELAY

This routine does nothing except use time. The amount of time used depends upon the
execution time of the various instructions. The maximum length of the delay is limited
by the size of the register; however. the entire routine can be placed Inside a similar
routine that uses another register. and so on.
The following example uses Register C and the Accumulator to
TRANSPARENT
provide delays as long as 255 ms. The chOice of registers is arDELAY
bitrary. You may. in fact. find the use of a register pair (e.g.. BC)
ROUTINE
more convenient. A PUSH BC instruction at the start of the
delay routine and a POP BC at the end will result In a routine that does not affect any
registers at all. Such a routine IS said to be "transparent" to the calling program. Note
that the PUSH and POP instructions must be included in the time budget

11-8

EXAMPLE
Delay Program Using Accumulator
Purpose: The program provides a delay of 1 ms times the contents of Accumulator.

Flowchart:

Count = MSCNT

Count

= Count -

(A) =(A)- 1

No

The value of MSCNT depends on the speed of the CPU and the memory cycle.

11-9

Source Program:
DELAY:
DLY1'

LD
DEC
JR
DEC
JR
RET

:GET COUNT FOR 1 MS DELAY

:COUNT = COUNT -1
:CONTINUE UNTIL COUNT = ZERO
:DECREMENT NUMBER OF REMAINING MS
:CONTINUE UNTIL NUMBER OF MS = ZERO

C.MSCNT

C
NZ.DLYl

A
NZ.DELAY

Object Program:
Memorv Location
(Hex)

Memorv Contents
(Hex)

0030
0031
0032
0033
0034
0035
0036
0037
0038

OE
MSCNT
00
20
FD
3D
20
F8
C9

Instruction
(Mnemonic)
DELAY:

LD

C.MSCNT

DLY1:

DEC
JR

C
NZ.DLYl

DEC
JR

A
NZ.DELAY

RET

Time Budget:
Instruction
LD
DEC
JR
DEC
JR
RET

Number of Times Executed

C.MSCNT

(A)

(A) x MSCNT
(A) x MSCNT
(A)
(A)

NZ.DLYl

A
NZ.DELAY

The total time used should be (A) x 1 ms. If the memory IS operating at full speed. the
Instructions require the following numbers of clock cYcles.
LD
DEC
JR
RET

C.MSCNT
C or DEC A
NZ

7
4
70r12
10

The alternative times for JR are for the condition being met (12) or not met (7).
Ignoring the CALL and RET instructions (which occur only once). the program takes:
(A) x (7+16 x MSCNT - 5 + 16) - 5
clock cycles. The -5's are caused by the fact that JR takes less time during the final
Iteration when the condition is not met.
So. to make the delay 1 ms.
13 + 16 x MSCNT = Nc
where Nc IS the number of clock cycles per millisecond. At the
standard 4 MHz Z80 clock rate. Nc = 4000. so:
16 x MSCNT = 3987
IMSCNT - 249 (hex F9) at a Z80 clock rate of 4 MHzl

11-10

zao DELAY
LOOP
CONS'rANT

SIMPLE I/O DEVICES

THE

zeo PARALLEL INPUT/OUTPUT CIRCUIT

(PIO)

The key element in most Z80 Input/output sections IS the Z80 Parallel Input/Output Circuit or PIO. This device combines latches, buffers, flip-flops, and other logic circuits
needed for handshaking and other simple interfacing techniques. The PIO contains
many logic connections, certain sets of which can be selected according to the contents of programmable registers. Thus, the deSigner has the eqUivalent of a Circuit
DeSigner's Casebook under his control. The initialization phase of the program places
the appropriate values into registers to select the required logic connections. An input/output section based on PIOs can handle many different applications, and changes
or corrections can be made in software rather than by rewIring.
Figure 11-7 is the block diagram of a PIO. The deVice contains two nearly identical 8-blt
ports - A, which IS usually an input port. and B, which IS usually an output port. Each
port (see Figure 11-8) contains:
An 8-bit Data Output register
PIO REGISTERS
AND
An 8-bit Data Input register
CONTROL LINES
A 2-bit Mode Control register, which indicates whether the
port IS In an output. input. bidirectional. or control mode
An 8-bit Input/Output Control register, which determines whether the corresponding
data pins are inputs (1) or outputs (0) in the control mode
Two control lines (STB and ROY) that are configured by the Mods Control register.
These lines can be used for the handshaking signals shown in Figures 11-5 and 11-6.
A 2-bit Mask Control register (used only in the control mode) that determines the active polarity of the inputs and whether they will be logically ORed or ANDed to form
an interrupt signal
An 8-bit Mask register (used only in the control model that determines which port
lines will be monitored to form the interrupt signal
An 8-bit Vector Address register used with the interrupt system
For now, we will be concerned only with the Mode Control registers, the Input/Output
Control registers, and the control lines. We will discuss the interrupt-related features of
the PIO in Chapter 12.
The meanings of the bits In the various control and mask registers are related to the underlying hardware and are entirely arbitrary as far as the assembly language programmer is concerned. You must either memorize them or look them up In this chapter and
in Chapter 12.
Each PIO occupies four input port addresses and four output port
addresses. The B/A SEL (Port B or A select) and C/O SEL (Control
or Data Select) lines choose one of the four ports as described in
Table 11-1. Most often, designers attach address bit AO to the B/A SEL input and address bit A1 to the C/O SEL input. The PIO then occupies four consecutive port addresses as described in the last column of Table 11-1.
Clearlv there are far more internal control registers than there are port addresses for
them, In fact. all the control registers for each port occupy one address according to the
C/O SEL connection. So some of the data bits sent to a control register are actually used
for addressing purposes. Note the following situations (see Table 11-2):
DO = 0 means that the remaining data bits are loaded into the Interrupt Vector register.

11-11

<I>

+5V GND

~~~

Penpheral
Interface
~

- , :.-

Intemal
Control

CPU
~

Data

I/O

L_

Internal

8us

CPU

......-

"

}HandShake

,_.

6
J

--

'---

8us

I/O

Control

Port A

r-V

8us

Pia
Control

,............,

Logic

Interface

Data or

Data or
Control

lines

Port 8

Interrupt
Control

Interrupt

Control lines

Figure 11-7 PIO Block Diagram

(Courtesy of Zilog)

11-12

I/O
~

} Handshake

Mode

Control Reg
(2 Bits!

Input/Output
Select Reg

(8 Bitsl

)?
Intemal8us

L--

.AMask
Control

-"

Reg

"V

12 Bitsl

Reg

(8 Bits!

Data Input
Input

(2 Bitsl

Data

I--

Penpheral

..,) Data or

Control Bus

Mesk
Reg

8-Bit

...

Data Output

Reg

(8 Bits!

<::-.

READ\

Interrupt Requests ....1---1

Handshake
Control
Logic

Figure 11-8. Block Diagram of PIO Port

(Courtesy of Zilogl

11-13

Handshake
}

14--

Lines

Table 11-1. PIO Addresses

CONTROL OR
DATA SELECT

PORT B OR A
SELECT

REGISTER
ADDRESSED

PORT ADDRESS
(STARTING WITH PIOADD)

0
0
1
1

0
1

Data Register A
Data Register B
Control A

PIOADD
PIOADD+1
PIOADD+2
PIOADD+3

0
1

Control B

The port addresses assume that C/D SEL IS tied to Al and B/A SEL to AO.

Table 11-2. Addressing of PIO Control Registers

REGISTER

ADDRESSING

MODE CONTROL
INPUT/OUTPUT CONTROL

D3 = D2 = D1 = DO = 1
NEXT WORD AFTER MODE CONTROL
SETS MODE 3

MASK CONTROL REGISTER

INTERRUPT MASK REGISTER

D3 = 0, D2 = 01 = DO = 1
NEXT WORD AFTER MASK CONTROL
REGISTER ACCESSED WITH D4 = 1

INTERRUPT ENABLE

D3=02=0, D1 =DO=1

INTERRUPT VECTOR

00= 1

03 = 0, D2 = D1 = DO = 1 means that the remaining data bits are loaded into the Mask
Control register. If D4 = 1, the next control word is loaded into the Interrupt Mask
register. Interrupts can be enabled or disabled with D3 = 02 = 0, D1 = DO = 1.
D3 = D2 = D1 = DO = 1 means that the remaining data bits are loaded into the Mode
Control register. If 07 = 06 1 (control model. the next control word is loaded into the
Input/Output Control register.

This sharing of an external address means that:

1)

2)

The programmer must be very careful of the order of operations. The meaning of a
particular Output instruction depends on the sequence in which it occurs.
The programmer should document the PIO configuration in detail. The device is
complex, and a reader is unlikely to be able to make much sense out of the sequence of operations that configures it.

We should note that one usually configures the control registers of the PIO just once in
the initialization phase of the program. The rest of the program then uses only the PIO
data registers.

11-14

PIO MODe CONTROL

(Pi'O'\
The mode of operation of a PIO is established by writing a control
word to the PIO in the form shown in Figure 11-1. Table 11-3 de~
scribes the meanings of the various modes and the control words
required to establish them. Note that bits 05 and 04 are not used. When power IS
turned on. the PIO comes up in mode 1 (input).
We may summarize the modes as follows:
1)

Mode 0 -

PIO

OUTPUT

OUTPUT
Writing data Into the port Output register latches the data and
MODE
causes it to appear on the port Data Bus. The READY (ROY)
line goes high to indicate Data Ready: it remains high until the
peripheral sends a rising edge on the STROBE (STB) line to indicate Data Accepted
or Device Ready. The rising edge of STB causes an interrupt if the Interrupt has
been enabled.
2)

Mode 1 -INPUT
PIO
INPUT
The peripheral latches data into the port Input register using
MODE
the STROBE signal. The rising edge of STB causes an Interrupt
(if enabled) and deactivates ROY. When the CPU reads the
data. ROY goes high to indicate Data Accepted or Input Register Empty. Note that
the peripheral can strobe data into the register regardless of the state of ROY. The
programmer must thus handle the problem of overrun. Le.. new data being placed
Into the register before the old data IS read.

3)

Mode 2 - BIDIRECTIONAL
PIO
BIDIRECTIONAL
This mode uses all four handshake lines. so It is allowed
MODE
only on Port A. The Port A ROY and STB signals are used
for output control and the Port B ROY and STB signals are
used for Input control. The only difference between this mode and a combination of
modes 0 and 1 is that data from the Port A Output register IS enabled onto the port
Data Bus only when A STB is active. This allows the Port A bus to be used bidirectionally under the control of A STB (Output Data Request) and B STB (Input Data
Available). Note that the B side control signals are governed bY Input Register A in
this mode.

4)

Mode 3 - CONTROL
This mode does not use the ROY and STB signals. It is intended for status and control applications in which each
bit has an individual meaning. When mode 3 IS selected.
the next control word sent to the PIO defines the directions
of the port data bits (Figure 11-9l. A '1' in a bit position
makes the corresponding bus line an input. while a '0'
makes it an output.

11-15

PIO
CONTROL
MODE
PIO
DIRECTIONS IN
CONTROL MODE

Set Mode

Ml

MO

Mode
Output

Input

1
1

0
1

Bidirectional
Bit Control

When selecting Mode 3. the next word must

set the I/O Register:

~
I/O = 1 Sets bit to Input
I/O = 0 Sets bit to Output

PIO Mode

Meening

.Control Word
(Binaryl (Hexl

Output

00001111

Input
Bidirectional

01001111

Control

11001111

2
3

10001111

OF
4F
BF
CF

Note that bits 4 and 5 are not used and could

have any values.

Figure 11-9. Mode Control for the ZSO PIO

Note the following features of the PIO modes:
1)

2)

3)
4)
5)

6)

FEATURES OF
PIO MODES
In modes 0.1, and 2 the peripheral indicates Data Ready.
DeVice Ready. or Data Accepted with a rising edge on the
STB line. This edge also causes an interrupt if the Interrupt IS enabled.
In modes 0.1. and 2 the PIO indicates Data Ready. Input Buffer Empty. or Data Accepted by sending ROY high. This signal remains high until the next rising edge on
STB.
Only Port A can be used bidirectionally. If Port A IS In mode 2 (bidirectional), Port B
can only be in mode 3 (cantrall since no handshake lines are available.
The control mode (3) IS the only mode In which the Input/Output Control register IS
used. Otherwise. the entire port IS used for either Input or output.
There is no way for the processor to determine if a pulse has occurred on STB if interrupts are not being used. The PIO is designed for use In interrupt-driven rather
than polling systems (see Chapter 12). STB should be tied low if it is not being
used.
The processor cannot directly control the ROY lines. The ROY line on a port goes
!:!.!@1 when data is transferred to or from the port and goes Iowan the rising edge of
STB.

7)

The contents of the data Output register can be read if the port is in the output or
bidirectional mode. If the port IS In the control mode. the output register data from
the lines assigned as outputs can be read. The contents of control registers cannot
be read.

S)

If the ROY output IS tied to the STB input on a port in the output mode. ROY will go
high for one clock period after each output operation. This brief pulse can be used
to multiplex displays as shown In Figure 11-1.

11-16

CONFIGURING THE PIO

The program must select the logic connections In the PIO before transfernng data to or
from it. This selection (or configuration) is usually part of the startup routine. Note that
the PIO comes up in the Input mode with all interrupts disabled and Inhibited and control signals deactivated (low) when power IS turned on. However. the PIO does not have
a RESET Input and does not necessarily return to the reset state when the CPU IS reset.
The steps in PIO configuration are:
1)

2)

STEPS IN
Establish the mode of operation by writing the apCONFIGURING A
propriate control words to the Mode Control register.
PIO
Interrupt control as well as I/O mode Information may
have to be sent.
If In mode 3. establish the directions of the I/O pins by writing a control word to the
Input/Output Control register. This word must follow the control word that selected
mode 3.

Let us now look at some examples of configuring a PIO without Interrupts:

1)

2)

3)

OUTPUT PORT
LD
A.00001111 B
OUT
(PIOCRBl.A

:MAKE PORT B OUTPUT

INPUT PORT
LD
OUT

:MAKE PORT A INPUT

A.01 001111 B
(PIOCRAl.A

BIDIRECTIONAL PORT
LD
A.10001111 B
OUT
(PIOCRAl.A

:MAKE PORT A BIDIRECTIONAL

Remember that only Port A can be bidirectional and that Port B must then be a control
port.
4)

CONTROL PORT. ALL INPUTS

LD
A.11001111 B
OUT
(PIOCRAl.A
LD
A.OFFH
OUT
(PIOCRAl.A

:MAKE PORT A CONTROL

:ALL BITS INPUTS

5)

CONTROL PORT. ALL OUTPUTS

LD
A.11001111 B
:MAKE PORT B CONTROL
OUT
(PIOCRBl.A
SUB
A
:ALL BITS OUTPUTS
OUT
(PIOCRBl.A

6)

CONTROL PORT. LINES 1.5.6 INPUTS; LINES 0.2.3.4. 7 OUTPUTS

LD
A.11 001111 B
:MAKE PORT A CONTROL
OUT
(PIOCRAl.A
LD
A.01100010B
:L1NES 1.5.6. INPUTS -0.2.3.4.7 OUTPUTS
OUT
(PIOCRAl.A

11-17

zao INPUT/OUTPUT INSTRUCTIONS

ZSO 1/0
The Z80 microprocessor has an extensive set of Input/Output
INSTRUCTIONS
instructions. All I/O instructions use 8-bit device addresses.
thus allowing up to 256 input ports and 256 output ports. But
remember that each PIO occupies four output port addresses and four input port addresses.
The I/O instructions can be grouped as follows:
1)

2)

3)

4)

Instructions that use absolute addressing. IN A. (portl and OUT (portlA transfer
eight bits of data between the Accumulator and the port addressed by the second
byte of the instruction.
Single-byte instructions that use register Indirect addressing. IN reg. (C) and OUT
(Cl,reg transfer eight bits of data between the specified register and the port addressed by Register C.
Block I/O instructions. INI and OUTI transfer eight bits of data between the memory
location addressed by Register Pair HL and the port addressed by Register C. Both
instructions then increment Register Pair HL and decrement the byte counter in
Register B. The Z flag IS set if B IS decremented to zero and reset otherwise. INO and
OUTO are the same instructions except that they decrement Register Pair HL instead of Incrementing it.
Repeated.Block I/O instructions. INIR and OTIR repeat the effects of INI and OUTI.
respectively. until B IS decremented to zero. INOR and OTOR have the same relationship to INO and OUTO.

You should note the following features of each group of instructions:

1)

Instructions with absolute addressing.

Data is always transferred to or from the Accumulator.
No flags are affected.
The port address is part of the program memory and
cannot be changed if that memory is read-only.

2)

Single-byte instructions with register indirect addressing.

Data can be transferred to or from any of the primary 8bit registers (A.B.C.O.E.H.U. However. remember that
Register C contains the port address.

1':1:":/0::::------.,

~~~R~~~~L~~
ADDRESSING

1/0
INSTRUCTIONS
WITH INDIRECT
ADDRESSING

IN reg.(C) sets the Sign (S). Zero (Zl. and Parity (P/O)
flags according to the value of the Input data. The Carry flag (C) is not modified.
but the Half Carry (H) and Negative (N) flags are reset. OUT (Cl,reg does not
affect any flags.

The port address is always in Register C. This address is not

11/0 DRIVER
part of the program memory and cou Id be a parameter for
an I/O subroutine (or I/O driver)' One I/O driver could thus be used In several
different applications or with several similar I/O devices in the same application.

11-18

3)

Block I/O instructions.

BLOCK I/O

Data is always transferred to or from the memory location INSTRUCTIONS

addressed by Register Pair HL.
The Z (Zero) flag is set if Register B is decremented to zero and cleared otherwise. The S (Sign), P/O (Panty)' and H (Half Carry) flags are affected. but their
final values are uncertain.
The port address IS always In Register C. Here again. this address could be a
parameter for an I/O dnver.
Register B IS an B-blt counter. Thus. the repeated Block I/O instructions can
transfer a maximum of 256 bytes. This differs from the Block Move and Block
Compare instructions. which use Register Pair BC as a 16-blt counter and can
handle up to 65K bytes.
Some examples of the various I/O instructions (without any
timing considerations) are:
1)

Load the Accumulator from Input Port 2.

a. USing absolute addressing
IN
A.(2)
b. Using register indirect addreSSing
LD
C.2
IN
A.(C)

2)

Store the contents of the Accumulator In Output Port 5.

a. USing absolute addreSSing
OUT
(5),A
b. Using register Indirect addreSSing
LD
C.5
OUT
(C),A

3)

Load memory location 0040 from Input Port 2.

a. Using absolute addressing
IN
A.(2)
:GET DATA
LD
(40H),A :STORE DATA
b. Using register indirect addressing
LD
C.2
:GET PORT NUMBER
IN
A.(C)
:GET DATA
LD
(40H),A ;STORE DATA
c. USing block I/O
LD
LD
INI

C.2
HLAOH

:GET PORT NUMBER

:GET MEMORY DESTINATION
:GET DATA

11-19

I/O
INSTRUCTION
EXAMPLES

4)

Store the contents of memory location 0040 in OutPUt t'ort 5.

a. Using absolute addressing
LD
A.(40H) :GET DATA
OUT
(5), A
:SEND DATA
b. Using register indirect addressing
LD
C.5
:GET PORT NUMBER
LD
A, (40H) :GET DATA
OUT
(C),A

c.

Using block

110

LD
LD
OUTI
5)

C.5
HLAOH

:GET PORT NUMBER

;GET MEMORY SOURCE
;SEND DATA

Load memory locations 0040 through 0047 from Input Port 2.

a. Using absolute addressing
LD
HLAOH ;GET STARTING ADDRESS OF DATA
;BYTE COUNTER = 8
LD
B.8
INBYTE: IN
A,(2)
;FETCH DATA BYTE
LD
(HL),A
;STORE BYTE IN MEMORY
INC
HL
DJNZ
INBYTE
b. Using block I/O

INBYTE:
C.

LD
LD
LD
INI
JR

HLAOH

B.8
C,2

;GET STARTING ADDRESS OF DATA

;BYTE COUNTER = 8
:GET PORT NUMBER

NZ,INBYTE

USing repeated block I/O

LD

HL,40H

lD

B,8

LD
INIR

C,2

;GET STARTING ADDRESS OF DATA

;BYTE COUNTER = 8
:GET PORT NUMBER
:MOVE INPUT BYTES TO MEMORY

1120

6)

Send the contents of memory locations 0040 through 0047 to Output Port 5.
a. USing absolute addressing
LD
HL40H
:GET STARTING ADDRESS OF DATA
LD
B.8
:GET BYTE COUNTER
OTBYTE: LD
A.(HL)
:FETCH BYTE FROM MEMORY
OUT
(51.A
:OUTPUT BYTE
INC
HL
DJNZ
OTBYTE
b. Using block I/O
LD
HLAOH
:GET STARTING ADDRESS OF DATA
B.8
:GET BYTE COUNTER
LD
LD
C.5
:GET PORT NUMBER
OTBYTE: OUTI
:OUTPUT BYTE FROM MEMORY
NZ.OTBYTE
JR
C.

USing repeated block I/O

LD
HLAOH
LD
B.8
LD
C.5
OTIR

:GET STARTING ADDRESS OF DATA

:GET BYTE COUNTER
:GET PORT NUMBER
:OUTPUT BYTES FROM MEMORY
~=~~~~
Note that the repeated Block I/O instructions operate conUSING BLOCK
tlnuously. You cannot provide any timing between
I/O INSTRUCTIONS
transfers. Thus. these Instructions cannot be used unless
the peripheral operates at the same speed as the processor or timing is handled separately in hardware. Ways to handle timing In hardware include forcing the processor
Into Walt states or buffering the data. Note that the Block I/O instructions all place the
contents of the byte counter (Register B) on the top half of the Address Bus during the
actual I/O transfer. In output operations. Register B is decremented first. The byte
counter value IS then available to external circuitry.
An obvious application for Block I/O instructions is the configuration of PIOs. Several
words must often be sent to a control register to determine operating mode. select pin
directions. and establish the interrupt system. No timing problems occur. since PIOs
operate at the same speed as the CPU. We will discuss the configuration of Z80 PIOs
and serial Interfaces (SIOs) With Block I/O instructions later In this chapter and In
Chapter 12.
In subsequent I/O examples. we will use mainly the instructions with absolute addressIng. You can easily substitute the instructions with register indirect addreSSing as long
as you remember to initialize Register C. We will occasionally Indicate applications for
the Block I/O instructions.

11-21

+5V

0
.1>
'1>
To CPU

"-

PIO
r

Pushbutton

-1

()

Figure 11-10. A Pushbutton Circuit

EXAMPLES
A Pushbutton Switch
Purpose: To Interface a single pushbutton switch (or a single-pole. single-throw (SPST)
sWitch) to a Z80 microprocessor. The pushbutton IS a mechanical switch that
provides a single contact closure (i.e.. a logic zero) while pressed.
Circuit Diagram:
Figure
a Z80
closed
pullup

11-10 shows the circuitry required to interface the pushbutton. It uses one bit of
PIO that acts as a buffer: no latch is needed. since the pushbutton remains
for many CPU clock cycles. Pressing the button grounds the PIO Input bit The
resistor ensures that the input bit is one if the button IS not being pressed.

Programming Examples:
We will perform two tasks With this circuit They are:
a)

Set a memory location based on the state of the button.

b)

Count the number of times that the button is pressed.

Task 1: Determine sWitch closure.

Purpose: Set memory location 0040 to one if the button IS not being pressed. and to
zero if It IS being pressed.
Sample Clises:
1)

Button open (i.e.. not pressed)

Result = (0040) = 01

2)

Button closed (i.e. pressed)

Result

= (0040) = 00

11-22

Flowchart:

(0040) = 0

Input and mask

pushbutton

data

(00401 = 1

Source Program:
LD
OUT
LD
LD
IN
AND

JR
DONE:

INC
HALT

A.Ol00llllB

:MAKE PORT A INPUT

(PIOCRAl.A

HL.40H
(HU,O

:MARKER =0

A.(PIODRA)
MASK
Z,DONE

:READ BUTTON POSITION

;IS BUTTON CLOSED (O)?
;YES, DONE
;NO. MARKER = 1

(HU

11-23

Object Program:

Memory Location
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010

3E
4F
03
PIOCRA
21
40
00
36
00
DB
PIODRA
E6
MASK
28
01
34
76

Instruction
(Mnemonic)
LD

A.Ol00llllB

OUT

(PIOCRA).A

LD

HL,40H

LD

(HU,O

IN

A.(PIODRA)

AND

MASK

JR

Z,DONE

INC
HALT

(HU

The port addresses PIOCRA and PIODRA depend on how the PIO is connected in your
microcomputer. The PIO control lines are not used In this example. In fact. we could
place the A side of the PIO In the control mode With the starting sequence:
LD
OUT
LD
OUT

A.ll00l111B
(PIOCRA).A
A.OFFH
(PIOCRA).A

:MAKEPORT A CONTROL
:ALL BITS INPUTS

MASK depends on the bit to which the pushbutton

button position and zeros elsewhere.

IS

connected; It has a one In the

Mask

Button Position
(Bit Number)

Binary

Hex

0
1
2
3
4
5
6
7

00000001
00000010
00000100
00001000
00010000
00100000
01000000
10000000

01
02
04
08
10
20
40
80

If the button is attached to bit 0 or bit 7 of the Input port. the program can use a Shift
instruction to set the Carry and thereby determine the button's state. For example.
Bit 7
IN
RLA
JR
Bit 0
IN
RRA
JR

A.(PIODRA)
NC.DONE
A,(PIODRA}
NC,DONE

:READ BUnON POSITION

:IS BUnON CLOSED (ZERO)?
:YES. DONE
:READ BUnON POSITION
;IS BUnON CLOSED (ZERO)?
:YES. DONE

11-24

The procedure for bit 7 is even simpler if we have the address of the PIO data register in
Register C. This is because the Input instructions using register indirect addressing
(e.g.. IN A,(C)) affect the Sign flag. The required sequence is:

..!.!l. (PIODRA
IN
JP

in Register C)
A,(C)
:READ BUTTON POSITION
P,DONE
:DONE IF BUTTON CLOSED (ZERO)

If the button is attached to bits 6 or 7 of the input port. the program can use the Sign bit
to determine the button's state. For example.
Bit 7
IN
AND
JP

A, (PIODRAl
A
P.DONE

:READ BUTTON POSITION

:IS BUTTON CLOSED (ZERO)?
;YES. DONE

IN A,(port} does not affect the flags: therefore, we must use the AND A instruction to
set the flags without changing the Accumulator.
Bit 6
IN
ADD
JP

A,(PIODRA)
A.A
P.DONE

:READ BUTTON POSITION

:IS BUTTON CLOSED (ZERO)?
:YES, DONE

RLA cannot be used because it does not affect the Sign bit.

11-25

Task 2: Count switch closures.

Purpose: Count the number of button closures by incrementing memory location 0040
after each closu reo
Sample Case:
Pressing the button ten times after the start of the program should give
(0040) = OA
Note: In order to count the number of times that the button has
been pressed. we must be sure that each closure causes a single
transition. However. a mechanical pushbutton does not produce a
single transition for each closure. because the mechanical contacts bounce back and
forth before settling Into their final positions. We can use a one-shot to eliminate the
bounce or we can handle It In software.

DEBOUNCING
The program can debounce the pushbutton by waiting after it
finds a closure. The reqUired delay IS called the debouncing
IN SOFTWARE
time and IS part of the specifications of the pushbutton. It IS
tYPically a few milliseconds long. The program should not examine the pushbutton durIng this period because It might mistake the bounces for new closures. The program
may either enter a delay routine like the one described previously or may simply perform other tasks for the specified amount of time.
Even after debouncing. the program must still wait for the present closure to end before
looking for a new closure. ThiS procedure avoids double counting. The following program uses a software delay of 1 ms to debounce the pushbutton. You may want to try
varYing the delay or eliminating It entirely to see what happens. To run thiS program.
you must also enter the delay subroutine into memory starting at location 0030.

Flowchart:

Count =0

No

Count

= Count

+ 1

Debounce button

with 1 ms walt

11-26

Source Program:

CHKCL:

CHKOP:

LD
OUT
LD
LD
IN
AND
JR
INC
CALL
IN
AND
JR
JR

A.Ol00llll B
(PIOCRAl.A
HLAOH
(HL).O
A. (PIODRA)
MASK
NZ.CHKCL
(HL)
DELAY
A.(PIODRA)
MASK
Z.CHKOP
CHKCL

:MAKE PORT A INPUT

:CLOSURE COUNT = ZERO

:READ BunON POSITION
:IS BUnON BEING PRESSED (D)?
:NO. WAIT UNTIL IT IS
;YES. INCREMENT CLOSURE COUNT
:WAIT 1 MS TO DEBOUNCE
:READ BUnON POSITION
:IS BUnON STILL BEING PRESSED (D)?
;YES. WAIT FOR RELEASE
:NO. LOOK FOR NEXT CLOSURE

Object Program:
Memory Location
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
DODD
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A

3E
4F
D3
PIOCRA
21
40
00
36
00
DB
PIODRA
E6
MASK
20
FA
34
CD
30
00
DB
PIODRA
E6
MASK
28
FA
18
EE

Instruction
(Mnemonic)

CHKCL:

CHKOP;

LD

A.Ol00lll1B

OUT

(PIOCRAl.A

LD

HLAOH

LD

(HL).O

IN

A.(PIODRA)

AND

MASK

JR

NZ.CHKCL

INC
CALL

(HU
DELAY

IN

A.(PIODRA)

AND

MASK

JR

Z.CHKOP

JR

CHKCL

The three instructions beginning with the label CHKOP are used to determine when the
sWitch reopens.
Clearly we do not really need a PIO for this simple interface. An addressable tn-state
buffer wou Id do the Job at far lower cost.

11-27

A Toggle Switch
Purpose: To interface a single-pole. double-throw (SPOT) toggle switch to a Z80
microprocessor. The toggle IS a mechanical device that is either in the normally closed (NC) position or the normally open (NO) position.
Circuit Diagram:
DEBOUNCING
Figure 11-11 shows the circuitry required to Interface the
WITH
switch. Like the pushbutton. the switch uses one bit of a Z80
PIO that serves as an addressable buffer. Unlike the button. the
CROSSCOUPLED
NAND GATES
switch may be left in either position. Typical program tasks are
to determine the switch position and to see if the position has
changed. Either a one-shot with a pulse length of a few milliseconds or a pair of crosscoupled NAND gates (see Figure 11-12) can debounce a mechanical switch.
The circuits will produce a single step or pulse In response to a change in switch position even if the switch bounces before settling into its new position.

Programming Examples:
We will perform two tasks involving this circuit. They are:
1)
2)

Set a memory location to one when the switch is closed.

Set a memorv location to one when the state of the sWitch changes.

Task 1: Wait for switch to close.

Purpose: Memorv location 0040 is zero until the switch is closed and then is set to
one: that is. the processor clears memory location 0040. walts for the sWitch
to be closed. and then sets memory location 0040 to one.
The sWitch could be marked Run/Halt. since the processor will not proceed until the
switch is closed.

Flowchart:

lOO4OJ = 0

(0040J

11-28

+5V

To CPU
Debounce

SWitch

Pia

Circuit

NO

Figure 11-11. A Toggle SWitch Circuit

+5V

To I/O Port (Pial

SWItch

-Figure 11-12. A Debounce Circuit Based on Cross-coupled NAND Gates

11-29

Source Program:

WAITC:

LD
OUT
LD
LD
IN
AND
JR
INC
HALT

A.01001111B
(PIOCRA),A
HLAOH
(HU.O
A.(PIOORA)
MASK
NZ.WAITC
(HU

:MAKE PORT A INPUT

:MARKER = ZERO
:READ SWITCH POSITION
;IS SWITCH CLOSED (ZERO)?
:NO. WAIT FOR SWITCH TO CLOSE
:YES. MARKER = 1

Object Program:
Memorv Location
(Hex)
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010

Instruction
(Mnemonic)

Memorv Contents
(Hex)
3E
4F
03
PIOCRA
21
40
00
36
00
DB
PIODRA
E6
MASK
20
FA
34
76

11-30

WAITC:

LD

A.01001111B

OUT

(PIOCRA),A

LD

HLAOH

LD

(HU,O

IN

A.(PIODRA)

AND

MASK

JR

NZ,WAITC

INC
HALT

(HU

Task 2: Wait for sWitch to change.

Purpose: Memory location 0040 remains zero until the sWitch position changes: I.e..
the processor walts until the sWitch changes. then sets memory location
0040 to 1.
Flowchart:

(401 = 0

Old data =

switch position

New data =
switch position

(401 = 1

Source Program:

SRCH:

LD
OUT
LD
LD
IN
AND
LD
IN
AND
CP

JR
INC
HALT

A.Ol00llllB
(PIOCRAl.A
HL,40H
(HL).O
A.(PIODRA)
MASK
B.A
A, (PIODRA)
MASK
B
Z.SRCH
(HL)

;MAKE PORT A INPUT

:MARKER = ZERO
:GET OLD SWITCH POSITION

:GET NEW SWITCH POSITION

;ARE NEW AND OLD POSITIONS THE SAME?
:YES. WAIT
:NO. MARKER = ONE

11-31

Object Program:
Memory Location
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
0008
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014

4F
03
PIOCRA
21
40
00
08
PIODRA
E6
MASK
47
08
PIOORA
E6
MASK
88
28
F9
34
76

Instruction
(Mnemonic!

3E

SRCH:

LD

A.01001111B

OUT

(PIOCRAl.A

LD

HL,40H

IN

A.(PIODRA)

AND

MASK

LD
IN

8.A
A.(PIOORA)

AND

MASK

CP
JR

8
Z.SRCH

INC
HALT

(HL)

A Subtract or Exclusive OR could replace the Compare in the program. Either of these
Instructions would. however. change the contents of the Accumulator. The Exclusive
OR would be useful if several SWitches were attached to the same Pia. since it would
produce a one bit for each switch that changed state. How would you rewrite thiS program so as to debounce the switch in software?

11-32

7
S
S
4

..

3
2
1

A7
AS
AS
A4
A3
A2
A1
AO

To CPU

PIO

Common

--Figure 11-13. A Multiple-Position SWitch

A Multiple-Position (Rotary, Selector, or Thumbwheell Switch

Purpose: To Interface a multiple-position sWitch to a microprocessor. The lead corresponding to the switch position IS grounded. while the other leads are high
(logic ones).
Circuit Diagram:

Figure 11-13 shows the circuitry required to Interface an 8-posltlon sWitch. The sWitch
uses all eight data bits of one side of a PIG. TYPical tasks are to determine the position
of the sWitch and to check whether or not that position has changed. Two special situations must be handled:
1)

The sWitch IS temporarily between positions so that no leads are grounded.

2)

The sWitch has not yet reached ItS final position.

The first of these situations can be handled by waiting until the input is not all ones. I.e..
until a switch lead is grounded. We can handle the second situation by examining the
switch again after a delay (such as 1 or 2 seconds) and only accepting the Input when
It remains the same. This delay will not affect the responsiveness of the system to the
switch. We can also use another SWitch (i.e.. a Load sWitch) to tell the processor when
the selector sWitch should be read.
Programming Examples:

We will perform two tasks involVing the circuit of Figure 11-13. These are:
a)
b)

Monitor the switch until it is In a definite position. then determine the posItion and
store its binary value in a memory location.
Walt for the position of the SWitch to change. then store the new position in a
memory location.

If the SWitch IS in a position. the lead from that position IS grounded through the common line. Pullup resistors on the input lines aVOid problems caused by noise.

11-33

Table 11-3. Data Input vs. Switch Position

Data Input

SWitch POSition
0
1

2
3
4

5
6
7

Binary

Hex

11111110
11111101
11111011
11110111
11101111
11011111
10111111
01111111

FE
FD
FB
F7
EF
DF
BF
7F

Task 1: Determine switch position.

Purpose: The program waits for the sWitch to be In a specific position and then places
the number of that posItion Into memory location 0040.
Table 11-3 contains the data Inputs corresponding to the various switch positions.
This scheme IS inefficient. since It requires eight bits to distinguish among eight
different posItIOns.
A TTL or MOS encoder could reduce the number of bits needed.
Figure 11-14 shows a cirCUit uSing the 74LS148 TTL 8-to-3 encoder 4 We attach the sWitch outputs in Inverse order. since the
74LS148 deVice has active-low inputs and outputs. The output of
the encoder cirCUit IS a 3-blt representation of the sWitch position. Many
clude encoders so that their outputs are coded. usually as a BCD digit
logic!.

USING
A TTL

ENCODER
switches in(in negative

The encoder produces active-low outputs. so. for example. switch position 5. which IS
attached to Input 2. produces an output of 2 in negative logic (or 5 In posItive logic!.
You may want to verify the double negative for yourself.

01
t

iQ

6
5
4

1,

12
13

3
2

'4

'is5

74LS146

8-10-3

~,

Encoder

00

17

::

A2
A,

To CPU

PIO

Ao

Common

"::?'
Figure 11-14. A Multiple-POSItIOn Switch with an Encoder

11-34

Flowchart:

Data

SWltch position

Position~O

Shift data
nght 1 bit

Yes

Position

=
(0040)

Positkln + 1

PositIon

Source Program:
LD
OUT
CHKSW: IN
CP
JR
LD
CHPOS. RRA
JR
INC
JR
DONE:
LD
LD
HALT

A,01001111B
(PIOCRAl.A
A,(PIODRA)
OFFH
Z,CHKSW
B,O
NC,DONE
B
CHPOS
HLAOH
(HL),B

;MAKE PORT A INPUT

:GET SWITCH DATA
;IS SWITCH IN A POSITION?
;NO, WAIT FOR A POSITION
:SWITCH POSITION = ZERO
;IS NEXT BIT GROUNDED POSITION?
;YES, SWITCH POSITION FOUND
;NO, INCREMENT SWITCH POSITION
;STORE SWITCH POSITION

11-35

Object Program:
Memorv Location
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016

3E
4F
03
PIOCRA
DB
PIODRA
FE
FF
28
FA
06
00
1F
30
03
04
18
FA
21
40
00
70
76

Instruction
(Mnemonic)
LD

A.01001111B

OUT

(PIOCRA).A

CHKSW: IN

CHPOS:

DONE:

A,(PIODRA)

CP

OFFH

JR

Z.CHKSW

LD

B.O

RRA
JR

NC.DONE

INC
JR

B
CHPOS

LD

HLAOH

LD
HALT

(HU.B

Suppose that a faulty switch or defective PIO results in the input always being OFF16.
How could you change the program so that it would detect this error?
There IS an unconditional jump. JR CHPOS. In the source program. Can vou change the
Initial conditions so as to make this Instruction unnecessary?
This example assumes that the sWitch is debounced in hardware. How would you
change the program to debounce the sWitch in software?

11-36

Task 2: Walt for switch position to change.

Purpose: The program waits for the sWitch position to change and places the new
position (decoded) into memory location 0040. The program walts until the
switch reaches its new position.

Flowchart:

Old data

Position

SWitch posi~ion

= -1

Shift data right 1 bit

New data

Position

=
Position + 1

SWitch position

(0040)

11-37

= PositIon

Source Program:

CHFST

LD
OUT
IN
CP

JR
CHSEC:

LD
IN
CP

JR
CP

JR
CHPOS.

LD
INC
RRA

JR
LD
LD
HALT

A,01001111B
(PIOCRA),A
A,(PIODRA)
OFFH
Z,CHFST
B,A
A,(PIODRA}
OFFH
Z.CHSEC
B
Z,CHSEC
B.OFFH
B
C,CHPOS
HL40H
(HLLB

:MAKE PORT A INPUT

:GET SWITCH DATA
:IS SWITCH IN A POSITION?
:NO, WAIT UNTIL IT IS
:GET NEW SWITCH DATA
:IS SWITCH IN A POSITION?
:NO, WAIT UNTIL IT IS
:IS POSITION SAME AS BEFORE?
:YES, WAIT FOR IT TO CHANGE
:SWITCH POSITION = -1
;INCREMENT SWITCH POSITION
:IS NEXT BIT GROUNDED POSITION?
:NO, KEEP LOOKING FOR GROUNDED POSITION
:STORE SWITCH POSITION

11-38

Object Program:
Memorv Location
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
DODD
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A
001B
001C
001D
DOlE

3E
4F
D3
PIOCRA
DB
PIODRA
FE
FF
28
FA
47
DB
PIODRA
FE
FF
28
FA
B8
28
F7
06
FF
04
IF
38
FC
21
40
00
70
76

Instruction
(Mnemonic)

CHFST

CHSEC:

CHPOS;

LD

A.Ol00llll B

OUT

(PIOCRAl.A

IN

A,(PIODRA)

CP

OFFH

JR

Z,CHFST

LD
IN

B.A
A,(PIODRA)

CP

OFFH

JR

Z,CHSEC

CP
JR

B
Z,CHSEC

LD

B.OFFH

INC
RRA
JR

B
C,CHPOS

LD

HLAOH

LD
HALT

(HU.B

An alternative method for determining if the switch is in a position is:

CHKSW: IN
INC
JR

A,(PIODRA)
A
Z,CHKSW

Why does this work? What happens to the input data?

11-39

A Single LED

zao

Purpose: To interface a single light-emitting diode to a

microprocessor. The LED
can be attached so that either a logic zero or a logic one turns it on.
Circuit Diagram:

Figure 11-15 shows the circuitry required to interface an LED. The

LED lights when its anode IS positive with respect to its cathode
(Figure 11-15a). Therefore, YOU can either light the LED by grounding the cathode and having the computer supply a one to the anode (Figure 11-15b) or
by connecting the anode to +5 volts and having the computer supply a zero to the
cathode (Figure 11-15cl. Using the cathode IS the most common approach. The LED is
brightest when It operates from pulsed currents of about 10 or 50 mA applied a few
hundred times per second. LEOs have a very short turn-on time (in the microsecond
range) so they are well sUited to multiplexing (operating several from a Single port). LED
circuits usually need peripheral or transistor dnvers and current-limiting resistors. MaS
devices normally cannot dnve LEOs directly and make them bnght enough for easy
viewing.
Note: The Pia has an output latch on each port. However. the B port is normally used
for output. since it has somewhat more drive capability. In particular. the B port outputs
are capable of driving Darlington transistors (providing 1.5 mA minimum at 1.5 V).
Darlington transistors are high-gain transistors capable of switching large amounts of
current at high speed: they are useful In driving solenoids, relays, and other deVices.
Task: Turn the light on or off.
Purpose: The program turns a single LED either on or off.

Send a LogiC One to the LED (turn a pOSitive display on or a negative display off).

Source Program:

(form data initially)

LD
OUT
LD
OUT
HALT

A.00001111 B
(PIOCRBl.A
A.MASKP
(PIODRB),A

:MAKE PORT B OUTPUT

:GET DATA FOR LED
:SEND DATA TO LED

An alternative uSing the control mode IS:

LD
A.11001111B
:MAKE PORT B CONTROL
OUT
(PIOCRBl.A
SUB
A
:MAKE ALL B LINES OUTPUTS
OUT
(PIOCRB),A
LD
A.MASKP
:GET DATA FOR LED
OUT
(PIODRBl.A
:SEND DATA TO LED
HALT
(update data)
IN
SET
OUT
HALT

A, (PIODRB)

LED,A
(PIODRBl.A

:GET OLD DATA

;TURN ON LED BIT
:SEND DATA TO LED

MASKP has a one bit In the LED position and zeros elsewhere. Note that we can read
the PIO Data Output register when the PIO IS In the output mode. We can also read any
combination of input data and output register data when the PIO is in the control
mode: the combination is defined by the assignment of Inputs and outputs.

11-40

A_nO_d-1e~lde

+svO~------------''VV'w

a)

BaSIC LED CircUItry, The resistor R should limit the maximum current to 50 rnA and

the average current to 10 mA.

From CPU

R
Dnver

PIO

hI

Interfacing an LED with positive logic. A iOglC T from the CPU turns the LED on.

~R

From CPU

PIO

cl

sv

Dnver

Interfacing an LED with negatlve logic. A logic '0' from the CPU turns the LEO on. The driver or the CPU
may Invert the logic levels.

Figure 11-15. Interfacing an LED

11-41

Object Program:
Memory Location
(Hex)

Memory Contents
(Hex)

Instruction
(Mnemonic)

(form data initially)

0000
0001
0002
0003
0004
0005
0006
0007
0008

3E
OF
03
PIOCRB
3E
MASKP
D3
PIODRB
76

LO

A.0000ll11 B

OUT

(PIOCRBl.A

LD

A.MASKP

OUT

(PIOORBl.A

HALT

(update data)
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
B.

OB
PIOORB
CB
LEO
D3
PIOORB
76

IN

A.(PIODRB)

SET

LEO.A

OUT

(PIOORB).A

HALT

Send a Logic Zero to the LEO !turn a positive display off or a negative displav on),
The differences are that MASKP must be replaced by its logical complement
MASKN and SET LEO.A must be replaced by RES LEO A Note that the second byte
of the object code for SET LED.A and RES LED.A depends on the actual bit position
to which the name LEO refers.
MASKN has a zero bit

In

the LED position and ones elsewhere.

11-42

B6

BS

f
e

B4

From. CPU

.J\.

PIO

-y

B3

Dnvers

Display

B2
B,

BO

Common

__1__
!CommonCathode I

87 may be used for decImal POlOt LED.

l6+sv

(CommonAnode)

Figure 11-16. Interfacing a Seven-Segment Display

Seven-Segment LED Display

Purpose: To Interface a seven-segment LED display to a Z80 microprocessor. The display may be either common-anode (negative logicl or common-cathode
(positive loglcl.
Circuit Diagram:
Figure 11-16 shows the circuitry required to Interface a
COMMON-ANODE
seven-segment display. Each segment may have one, two.
OR
or more LEOs attached in the same way. There are two
COMMON-CATHODE
ways of connecting the displays. One is tYing all the
DISPLAYS
cathodes together to ground (see Figure 11-17a); this is a
"common-cathode" display. TYing all the anodes together to a POSitive voltage supply
(see Figure 11-17bl is a "common-anode" display, and a logic zero at a cathode lights a
segment. So the common-cathode display uses positive logic and the common-anode
display negative logic. Either display requires appropriate drivers and resistors.
The Common line from the display
segments are customarily labelled:

IS

tied either to ground or to +5 volts. The display

a

9
e

11-43

a)

Common-cathode

go----..It/'I/Ir---tiIDlIt-----,

eo-----~"'"""--t.tt_----...,
d o-----"rw--~ti-----...

...

b o-----J\rw--~ti-----

oo------'\iM~--ti.H_----_...

-b)

Common-anode

+5V

go-----"M---ttllBI-----4j9

o------I\il\J\l---ti.H_------f

do-----J\rw--tfIIIltt-----...

...IH----__f

co----'tN~-

bo----....l\M---HllIli-------1
oo-----..I'/i~---tt.tt_------I

Figure 11-17, Seven-Segment Display Organization

11-44

Table 11-4. Seven-Segment Representations of Decimal Numbers

Number
0
1
2
3
4
5
6
7
8
9

Hexadecimal Rep.resentation
Common-anode
Common-cathode
3F
06
5B
4F
66
6D
7D
07
7F
67

40
79
24
30
19
12
02
78
00
18

Bit 7 is always zero and the others are g. f. e. d. c. b. and a in decreasing order of
slgnifi cance.
Note: The seven-segment display IS widely used because it
SEVEN-SEGMENT
contains the smallest number of separately controlled segREPRESENTATIONS
ments that can provide recognizable representations of all
the decimal digits (see Figure 11-18 and Table 11-4). Seven-segment displays can also
produce some letters and other characters (see Table 11-5). Better representations reqUire a substantially larger number of segments and more circuitry. 5 Since seven-segment displays are so popular. low-cost seven-segment decoder/drivers have become
widelY available. The most popular devices are the 7447 common-anode driver and the
7448 common-cathode dnver6 : these devices have Lamp Test Inputs (that turn all the
segments on) and blanking inputs and outputs (for blanking leading or trailing zeros).

11-45

0: Segments f. e, d. c, b. a

3: Segments g. d. c, b. a Q!J

1: Segments

C,

4: Segments 9. f,

C,

b 2!:1

5: Segments 9, f, d, C, a 2!)

2: Segments 9, e, d, b, a 2!)

I,

Figure 11-18. Seven-Segment Representations of Decimal Digits

11-46

6: Segments g, f, e, d. c. a

e!:'

8: Segments g. f. e, d. c, b. a

Note that the alternate representation with a off may

be reserved for the lower case letter 'b'. 7: Segments c. b, a

This is the same as LAMP TEST.

9: Segments g, f. c, b. a 2r'

An alternate has segment d on also.

Figure 11-18. Seven-Segment Representations of Decimal Digits

(Continued)

11-47

Table 11-5. Seven-Segment Representations of Letters and Symbols

Upper-case Letters
Letter

Hexadecimal Representation
Common-cathode

Common-anode

77

C
E
F
H
I

39
79
71
76
06
1E
38
3F
73
3E
66

08
46
06
OE
09
79
61
47
40
OC
41
19

J
L

0
p
U

Lower-case Letters and Special Characters

Character
b
c
d
h
n

a
r
u
7

Hexadecimal Representation
Common-cathode

Common-anode

7C
58
5E
74
54
5C
50
1C
40
53

03
27
21
OB
2B
23
2F
63
3F
2C

11-48

Task 1: Display a decimal digit.

Purpose: Display the contents of memory location 0040 on a seven-segment display if
it contains a decimal digit. Otherwise. blank the display.
Sample Problems:
a.

(0040) = 05

b.

Result IS 5 on display
(0040) = 66
Result is a blank display

Flowchart:

Code
Data

Blank
(0040)

Code = (SSEG +
Dala)

Send code

to display

11-49

Source Program:
LD
OUT
lD
lD
CP

JR
LD

DSPlY:

LD
LD
ADD
LD
lD
OUT
HALT

A.00001111 B
(PIOCRBl.A
B,BlANK
A,(40H)
10
NC,DSPlY
DE,SSEG
H,O
l,A
HL.DE
B,(HL)

;MAKE PORT B OUTPUT

;GET BLANK CODE
;GET DATA
:IS DATA A DECIMAL DIGIT?
:NO, DISPLAY BLANKS
:GET BASE ADDRESS OF SEVEN-SEGMENT
, TABLE
:MAKE DATA INTO 16-BIT INDEX
:ACCESS ELEMENT IN TABLE
:GET SEVEN-SEGMENT CODE

AB
(PIODRBl.A

:SEND CODE TO DISPLAY

BLANK is 00 for a common-cathode display, FF for a common-anode display, An alternative procedure would be to put the blank code at the end of the table and replace all
Improper data values with 10, I.e..
lD
CP

JR
CNVRT:

LD
LD

A,(40Hl
10
C,CNVRT
A,10
DE,SSEG

;GET DATA
;IS DATA A DECIMAL DIGIT?
:YES, CONVERT DIRECTLY TO SEVEN-SEGMENT
:NO, GET INDEX FOR BLANK CODE
:GET BASE ADDRESS OF SEVEN-SEGMENT TABLE

Table SSEG IS either the common-cathode or common-anode representation of the

decimal digits from Table 11-4.

11-50

Object Program:
Memory Location
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0020-0029

3E
OF
D3
PIOCRB
06
BLANK
3A
40
00
FE
OA
30
08
11
20
00
26
00
6F
19
46
78
D3
PIODRB
76

Instruction
(Mnemonic!

DSPLY'

SSEG:

LD

A.OOOOllll B

OUT

(PIOCRB),A

LD

B.BLANK

LD

A.(40H)

CP

10

JR

NC.DSPLY

LD

DE.SSEG

LD

H.O

LD
ADD
LD
LD
OUT

L.A
HL.DE
B.(HL)
A.B
(PIODRBl.A

HALT
(seven-segment code
table)

Several displays may be multiplexed. as shown In Figure 11-19. A brief strobe on the B
RDY line clocks the counter and directs data to the next display. Note that B RDY IS tied
directly back to B STB. Le.. the ready line essentially provides ItS own acknowledgment.
The timing of the PIO IS such that this connection results in a strobe with a duration of
one clock period. Such a brief strobe is exactly what the counter requires. RESET starts
the deCimal counter at nine so that the first output operation clears the counter and
directs data to the first display.
The follOWing program uses the delay routine to pulse each of ten common-cathode
displays for 1 ms.

11-51

O. C. B. and A (0 most significant. A

least significant) are the 4-bit output
from the counter. These 4 bits actiVate

the correspondingly numbered output

from the decoder. and hence the correspondingly numbered displav.

From CPU

PIO

B ROY

Oock
Decade

Counter
Reset

4 to 10
Decoder/Dover

B
A

Figure 11-19. Multiplexed Seven-Segment Displays

11-52

Task 2: Display ten decimal digits.

Purpose: Display the contents of memory locations 0040 through 0049 on ten 7-segment displays that are multiplexed with a counter and a decoder.
Sample Problem:

(0040)
66
(0041)
3F
(0042)
7F
(0043)
7F
(0044)
06
(0045)
5B
(0046)
07
(0047)
4F
(0048)
6D
(0049)
7D
Display reads 4088127356
Source Program:

DRUN:

DSPLY:

LD
OUT
LD
LD
LD
OUT1
CALL
JR
JR

A.00001111 B
(PIOCRB).A
HLAOH
B.10
C.PIODRB
DELAY
NZ.DSPLY
DRUN

;MAKE PORT B OUTPUT

;POINT TO START OF DATA
:NUMBER OF DISPLAYS = 10
;GET PORT NUMBER
;SEND DATA TO DISPLAY
:WAIT 1 MS
:COUNT DISPLAYS
;START ANOTHER SCAN

Here we must select the PIO output mode. since the circuit uses the handshake signals.
Note that OUTI sends the data to the output port addressed by Register C. Increments
the address in Register Pair HL. and decrements the counter in Register B. We have
assumed that subroutine DELAY does not affect the Z flag so that it can be used afterwards for a conditional branch.

11-53

Object Program:

Memorv Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013

3E
OF
03
PIOCRB
21
40
00
06
OA
OE
PIOORB

ED

Instruction
(Mnemonic)

ORUN:

DSPLY:

A3
CO
30
00
20
F9
18
FO

11-54

LO

A.00001111 B

OUT

(PIOCRBl.A

LO

HLAOH

LD

B.10

LD

C.PIOORB

OUTI
CALL

DELAY

JR

NZ.OSPLY

JR

ORUN

PROBLEMS
1) An On-Off Pushbutton
Purpose: Each closure of the pushbutton complements (inverts) all the bits in memory
location 0040. The location Initially contains zero. The program should continuouslY examine the pushbutton and complement location 0040 with each
closure. You may wish to complement a display output port instead. so as to
make the resu Its easier to see.
Sample Case:

Location 0040 initially contains zero.

The first pushbutton closure changes location 0040 to FF (hex). the second changes it
back to zero. the third back to FF (hex). etc. Assume that the pushbutton is debounced
in hardware. How would you Include debouncing In your program?

2)

Debouncing a Switch in Software

Purpose: Debounce a mechanical switch by waiting until two readings. taken a debounce time apart. give the same result Assume that the debounce time (in
ms) is in memory location 0040 and place the SWitch position into memory
location 0041.
Sample Problem:

(0040) = 03 causes the program to walt 3 ms between readings.

3)

Control for a Rotary Switch

Purpose: Another SWitch serves as a Load SWitch for a four-position unencoded rotarY
switch. The CPU walts for the Load switch to close (be zero). and then reads
the position of the rotarY SWitch. This procedure allows the operator to move
the rotary SWitch to ItS final position before the CPU tries to read It The program should place the pOSition of the rotarY SWitch into memory location
0040. Debounce the Load SWitch In software.
Sample Problem:

Place rotary switch in position 2. Close Load switch.

Result:

(0040) = 02

11-55

4)

Record Switch Positions on lights

Purpose: A set of eight sWitches should have their positions reflected In eight LEOs.
That is to say. if the switch is closed (zero). the LED should be on. otherwise
the LED should be off. Assume that the CPU output port is connected to the
cathodes of the LEOs.
Sample Problem:
SWITCH
SWITCH
SWITCH
SWITCH
SWITCH
SWITCH
SWITCH
SWITCH

CLOSED
OPEN
CLOSED
OPEN
OPEN
CLOSED
CLOSED
OPEN

ON

1
2
3
4

5
6

Result:
LED
LED
LED
LED
LED
LED
LED
LED

1 OFF

2 ON
3 OFF
4

OFF

5 ON
6 ON
7

OFF

How would you change the program so that a switch attached to bit 7 of Port A of PIO
#2 determines whether or not the displays are active (j.e., if the control switch is
closed. the displays attached to Port B reflect the switches attached to Port A: if the
control switch is open. the displays are always off)? A control switch is useful when the
displays may distract the operator. as In an airplane.
How would you change the program so as to make the control sWitch an on-off
pushbutton; that IS. each closure reverses the previous state of the displays? Assume
that the displays start In the active state and that the program examines and debounces
the pushbutton before sending data to the displays.

5)

Count on a Seven-Segment Display

Purpose: The program should count from 0 to 9 continuously on a seven-segment display. starting with zero.
Hint: Try different timing lengths for the displays and see What happens. When does
the count become visible? What happens if the display is blanked part of the time?

11-56

MORE COMPLEX I/O DEVICES

More complex I/O devices differ from simple keyboards. switches. and displays In that:
1)

They transfer data at higher rates.

2)

They may have their own Internal clocks and timing.

3)

They produce status Information and require control information. as well as

transferring data.

Because of their high data rates. you cannot handle these I/O devices casually. If the
processor does not provide the appropriate service. the system may miss Input data or
produce erroneous output data. You are therefore working under much more exacting
constraints than In dealing with simpler devices. Interrupts are a convenient method
for handling complex I/O devices. as we shall see in Chapter 12.
Peripherals such as keyboards. teletypewnters. cassettes.
SYNCHRONIZING
WITH I/O
and floppy disks produce their own Internal timing. These
DEVICES
devices provide streams of data. separated by specific timIng Intervals. The computer must synchronize the initial input or outPut operation with the penpheral clock and then provide the proper Interval
between subsequent operations. A Simple delay loop like the one shown previously can
produce the timing Interval. The synchronization may require one or more of the followIng procedures:
1)

Looking for a transition on a clock or strobe line provided by the penpheral for timIng purposes. A simple approach would be to tie the strobe to a PIO STB input and
look for a change in the Interrupt (INT) output. However. there is no way to directly
address the INT output (and thus determine ItS value) and no way to clear It other
than through an interrupt service routine. Thus. to use the PIO in a polling system.
one must make the strobe available at an Input port and latch It if necessary. If the
strobe must be latched. a circuit must also be provided to clear the latch as part of
the subsequent Input or output transfer.

2)

Finding the center of the time Interval during which the data is stable. We would
prefer to determine the value of the data at the center of the pulse rather than at
the edges. where the data may be changing. Finding the center requires a delay of
one-half of a transmission Interval (bit time) after the edge. Sampling the data at
the center also means that small timing errors have little effect on the accuracy of
the reception.

3)

Recognizing a special starting code. ThiS is easy if the code IS a single bit or if we
have some timing information. The procedure IS more complex if the code is long
and could start at any time. Shifting will be necessary to determine where the
transmitter is starting ItS bits. characters. or messages (this IS often called a search
for the correct "framing").

4)

Sampling the data several times. This reduces the probability of receiVing data incorrectly from nOIsy lines. MaJonty logic (such as best 3 out of 5 or 5 out of 8) can
be used to deCide on the actual data value.

Reception IS. of course. much more difficult than transmission. since the penpheral controls the receptIOn and the computer must Interpret timing information generated by
the penpheral. In transmiSSion. the computer provides the proper timing and formatting
for a specific penpheral.
Penpherals may require or provide other Information beSides
CONTROL
data and timing. We refer to other information transmitted by
AND STATUS
the computer as "control information". It may select modes of
INFORMATION
operatIOn. start or stop processes. clock registers. enable
buffers. choose formats or protocols. provide operator displays. count operations. or

11-57

Identify the type and prionty of the operation. We refer to other information transmitted
by the peripheral as "status information"; it may Indicate the mode of operation. the
readiness of devices. the presence of error conditions. the format of protocol in use, and
other states or conditions.
The computer handles control and status information just like data. This information
seldom changes, even though actual data may be transferred at a high rate. The control
or status Information may be single bits. digits. words, or multiple words. Often single
bits or short fields are combined and handled by a single input or output port.
Combining status and control information into bytes reduces the total number of I/O
port addresses required by the peripherals. However. the combination does mean that
Individual status Input bits must be separately interpreted and control output bits must
be separately determined. The procedures for Isolating status bits and setting or resetting control bits are as follows:
Separating Out Status Bits

SEPARATING
STATUS
INFORMATION

Step 1)

Read status data from the penpheral

Step 2)

Logical AND with a mask (the mask has ones In bit

positions that must be examined and zeros
elsewhere)

Step 3)

Shift the separated bits to the least significant bit positions

If the field IS a single bit. Step 2 IS unnecessary since we can test the bit With the BIT instruction. If the single bit is In the most significant. next to most significant. or least significant position. we can use shift logical (AND .A or OR A) instructions to determine its
value. Remember also that the Input instructions With register indirect addressing (e.g..
IN A.(C)) affect the Sign flag. These somewhat more accessible bit positions are often
reserved for the most frequently used status information. You Should try to write the reqUired Instruction sequences for the Z80 processor.
Step 3 is unnecessary if the field is a single bit. since the Zero flag will contain the complement of that bit after Step 2 (try It!). A Shift or Load instruction can replace Step 2 if
the field IS a single bit and occupies the least significant. most significant. or next to
most significant bit position. These POSitions are often reserved for the most frequently
used status information. You should try to wnte the required Instruction sequences for
the 6800 processor.
Setting and Clearing Control Bits
Step 1)
Step 2)
Step 3)
Step 4)

COMBINING
CONTROL
INFORMATION

Read pnor control Information

Logical AND with mask to clear bits (mask has zeros
in bit positions to be cleared, ones elsewhere)
Logically OR with mask to set bits (mask has ones in bit positions to be set.
zeros elsewhere)
Send new control information to peripheral

Here again the procedure is simpler if the field IS a single bit and occupies a position at
the end of the word.

11-58

Some examples of separating and combining status bits are:

1)

A 3-bit field In bit positions 2 through 4 of a PIO data register IS a scaling factor.
Place that factor Into the Accu mulator.

READ STATUS DATA FROM INPUT PORT

IN

A.(PIQDR)

:READ STATUS DATA

MASK OFF SCALING FACTOR AND SHIFT

AND
RRCA
RRCA
2)

00011100B

. ;MASK SCALING FACTOR

;SHIFT TWICE TO NORMALIZE

The Accumulator contains a 2-blt field that must be placed into bit positions 3 and
4 of a PIO data register.

,MOVE DATA TO FIELD POSITIONS

RLA
RLA
RLA
AND
LD

:SHIFT DATA TO BIT POSITIONS 3 AND 4

00011000B
B.A

:CLEAR OTHER BIT POSITIONS

:SAVE NEW FIELD VALUE

COMBINE NEW FIELD VALUE WITH OTHER DATA

IN
AND
OR
OUT

A.(PIODR)
11100111B
B
(PIODR),A

;CLEAR OLD FIELD VALUE

;INSERT NEW FIELD VALUE

Documentation IS a serious problem In handling control and

DOCUMENTING
status information. The meanings of status inputs or control
STATUS AND
outputs are seldom obvious. The programmer should clearly inCONTROL
dicate the purposes of Input and output operations In the comTRANSFERS
ments. e.g.. "CHECK IF READER IS ON:' "CHOOSE EVEN
PARITY OPTION:' or "ACTIVATE BIT RATE COUNTER." The bit manipulation. Logical.
and Shift instructions will otherwise be very difficult to remember. understand. or
debug.

11-59

Table 11-6. Comparison Between Independent Connections

and Matrix Connections for Keyboards
Keyboard Size

Number of Lines with

Independent Connections

Number of Lines with

Matrix Connections

6
8
10
10
12
14
16

3x3
4x4
4x6
5x5
6x6
6x8
8x8

16
24
25
36
48
64

EXAMPLES
An Unencoded Keyboard
Purpose: Recognize a key closure from an unencoded 3 x 3 keyboard and place the
number of the key that was pressed mto the Accumu lator.
Keyboards are Just collections of switches (see Figure 11-20). Small numbers of keys are
easiest to handle if each key is attached separately to a bit of an input port. Interfacing
the keyboard IS then the same as mterfacmg a set of sWitches.
Keyboards with more than eight keys require more than one input
port and therefore mu Itibyte operations. This IS partlcu larly
wasteful if the keys are logically separate. as in a calculator or terminal keyboard where the user will only strike one at a time. The number of mput lines
required may be reduced by connectmg the keys into a matrix. as shown in Figure
11-21. Now each key represents a potential connection between a row and a column.
The keyboard matrix requires n + m external lines. where n IS the number of rows and
m is the number of columns. This compares to n x m external lines if each key IS separate. Table 11-6 compares the number of keys required by typical configurations.
A program can determme which key has been pressed by usmg
KEYBOARD
the external lines from the matrix. The usual procedure is a
SCAN
"keYboard scan." We ground Row 0 and examine the column
lines. If any lines are grounded. a key in that row has been pressed. causmg a row-tocolumn connection. We can determine which key was pressed by determming which
column line is grounded: that is. which bit of the input port is zero. If no column line IS
grounded. we proceed to Row 1 and repeat the scan. Note that we can check to see if
any keys at all haye been pressed by grounding all the rows at once and examinmg the
columns.
The keyboard scan requires that the row lines be tied to an output port and the column
lines to an input port. Figure 11-22 shows the arrangement. The CPU can ground a particular row by placmg a zero m the appropriate bit of the output port and ones m the
other bits.
The CPU can determine the state of a particular column by examining the appropriate
bit of the input port.

11-60

_r
_r
_r

Kev 1

..L

Key 2

.1.

Kev3

..L

Each key IS a switch lust like a pushbutton and grounds an Input Qit if it is pressed.

...;-

Figure 11-20. A Small Keyboard

Column 0

Column 1

Column 2

Row 0

Row 1

Row 2
Each key now serves to connect a row to a column. For Instance. kay 4 connects row 1 to column 1.

Figure 11-21. A Keyboard Matrix

11-61

Column 0

Data Bus
(from CPU)

Column 1

Coiumn 2

00 t---If---+---4II------t--~IJ_--t--Row

0,1I---"---t---f1---+---4F---t-- Row

PIO
Output

Port

OzIl---.---t--......IIf---+---4If---t--Row 2

10

PIO
Input

Port

Data Bus (to CPU)

Figure 11-22. I!OArrangement for a Keyboard Scan

11-62

Task 1: Determine key closure.

Purpose: Wait for a key to be pressed.

WAITING

The procedure is as follows:

FORA
KEY CLOSURE

1l

Ground all the rows by cleanng all the output bits.

2)

Fetch the column Inputs by reading the Input port.

Return to Step 1 if all the column inputs are ones.

3)

Flowchart:

Ground all
keyboard rows

Source Program:

WAITK,

LD
OUT
LD
OUT
SUB
OUT
IN
AND
CP

JR

A.01001111B
(PIOCRA).A
A00001111 B
(PIOCRB),A
A
(PIODRB).A
A,(PIODRA)
00000111B
00000111B
Z,WAITK

:MAKE PORT A INPUT

:MAKE PORT B OUTPUT
:GROUND ALL KEYBOARD ROWS
:GET KEYBOARD COLUMN DATA
;MASK COLUMN BITS
:ARE ANY COLUMNS GROUNDED?
;NO, WAIT UNTIL ONE IS

HALT

11-63

Object Program:

Memorv Location
(Hex)

Memorv Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013

3E
4F
03
PIOCRA
3E
OF
03
PIOCRB
97
03
PIODRB
DB
PIODRA
E6
07
FE
07
28
F8
76

Instruction
(Mnemonic)

WAITK:

LD

A.01001111B

OUT

(PIOCRAl.A

LD

A.00001111 B

OUT

(PIOCRBl.A

SUB
OUT

A
(PIODRBl.A

IN

A.(PIODRA)

AND

00000111B

CP

00000111B

JR

Z,WAITK

HALT

PIO Port B IS the keyboard output port and Port A IS the Input pon.
Masking off the column bits eliminates any problems that could be caused by the states
of the unused Input lines.
We could generalize the routine by naming the output and masking patterns:
ALLG
EQU
11111000B
OPEN
EQU
00000111 B
These names could then be used In the actual program: a different keyboard would require only a change In the definitions and are-assembly.
Of course. one port of a PIO IS all that IS really necessary for a 3 x 3 or 4 x 4 keyboard.
Try rewriting the program so that It uses only Port A. The PIO must be placed into the
control mode so that lines can be indiVidually selected as inputs or outputs.

11-64

Task 2: Identify key.

Purpose: Identify a key closure by placing the number of the key into the Accumulator.

The procedure is as follows:

1)

Set key number to -1. counter to number of rows. and output pattern to all ones
except for a zero In bit O.

2)
3)
4)
5)
6)
7)

Ground a row by sending the output pattern to the keyboard output port.
Update the output. pattern by shifting the zero bit left one position.
Fetch the column inputs by reading the input port.
If any column inputs are zero. proceed to Step 8.
Add the number of columns to the key number to reach the next row.
Decrement counter. Go to Step 2 if any rows have not been scanned. otherwise to
Step 10.
Add 1 to key number. Shift columh inputs nght one bit.
If Carry = 1. return to Step 8.

8)
9)
10)

End of program.

Flowchart:

Ground row bv
output of

Scan Pattern

Kev Number =
Kev Number + 1
Shift column inputs
nght 1 bit

Update Scan Pattern

by shifting left
circularlv

Kev Number =
Kev Number +
Number of Columns
Counter = Counter - 1
Ves

11-65

Source Program:

FROW:

FCOL:

DONE,

LD
OUT
LD
OUT
LD
LD
LD
LD

A.01001111B
(PIOCRA).A
A.00001111 B
(PIOCRB).A
B.3
C.PIODRB
D.3
E. 1111111 OB

LD
LD
OUT
RLC
IN
AND
CP
JR
LD
ADD
LD
DJNZ
INC
JR
INC
RRA
JR
HALT

H.00000111 B
L.OFFH
(C),E
E
A.(PIODRA)
H
H
NZ.FCOL
A.L
A.D
L.A
FROW
L
DONE
L
NC.FCOL

:MAKE PORT A INPUT

:MAKE PORT B OUTPUT
:COUNT = NUMBER OF ROWS
:GET OUTPUT PORT NUMBER
:GET NUMBER OF COLUMNS
;START SCAN PATIERN TO GROUND ROW
ZERO
:GET KEYBOARD MASKING PATIERN
:KEY NUMBER = -1
:SCAN A ROW
:UPDATE SCAN PATIERN FOR NEXT ROW
:GET KEYBOARD COLUMN DATA
:MASK COLUMN BITS
:ARE ANY COLUMNS GROUNDED?
:YES. GO FIND WHICH ONE
:NO. UPDATE KEY NUMBER FOR NEXT ROW

;EXAMINE NEXT ROW IF ANY LEFT

;IDENTIFY CASE IN WHICH KEY NOT FOUND
;INCREMENT KEY NUMBER
;IS THIS COLUMN GROUNDED?
;NO. EXAMINE NEXT COLUMN

11-66

Object Program:
Memorv Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A
001B
001C
0010
001E
001F
0020
0021
0022
0023
0024
0025
0026
0027
0028
0029
002A

3E
4F
03
PIOCRA
3E
OF
03
PIOCRB
06
03
OE
PIODRB
16
03
1E
FE
26
07
2E
FF
ED
59
CB
03
DB
PIODRA

Instruction
(Mnemonic)

FROW'

A4
BC
20
08
70
82
6F
10
F1
2C
18
04
2C
1F
30
FC
76

FCOL:

LD

A.01001111B

OUT

(PIOCRAl.A

LD

A.00001111 B

OUT

(PIOCRBl.A

LD

B.3

LD

C.PIODRB

LD

0.3

LD

E.11111110B

LD

H.00000111 B

LD

L.OFFH

OUT

(Cl.E

RLC

IN

A.(PIODRA)

AND
CP
JR

H
H
NZ.FCOL

LD
ADD
LD
DJNZ

A.L
A.D
L.A
FROW

INC
JR

L
DONE

INC
RRA
JR

L
NC.FCOL

HALT

Each time a row scan fails. we must add the number of columns to the key number so
as to move past the present row (try It on the keyboard In Figure 11-22).

11-67

What is the result of the program if no keys are being pressed? Note the extra INC L instruction so that the program differentiates between no keys pressed and the last key
being pressed. What IS the final value in the Accumulator for these two cases? Note
that the Zero flag could also be used to distinguish the case where no keys were
pressed. Can you explain how?
An alternative approach would be to use the PIG in Its control mode so that lines could
be changed from Inputs to outputs. The procedure would be:
1)
2)

Ground all the columns and save the row Inputs.

Ground all the rows and save the column Inputs.

3)

Use the row and column Inputs together to determine the key number from a table.

Try to write a program to implement this procedure.

This program can be generalized by making the number of rows. the number of columns. and the masking pattern Into named parameters with EQU pseudo-operations.

11-68

':'0

~ Kevboard Data Inputs

A7
Data Bus
(
to CPU

PIO
ASTB

B4

.;J

Kevboard Strobe

(---FL --u- )
or

Figure 11-23. I/O Interface for an Encoded Keyboard

An Encoded Keyboard
Purpose: Fetch data. when it IS available. from an encoded keyboard that provides a
strobe along with each data transfer.

An encoded keyboard provides a unique code for each key. It has Internal electronics
that perform the scanning and Identification procedure of the prevIous example. The
tradeoff IS between the simpler software required by the encoded keyboard and the
lower cost of the unencoded keyboard.
Encoded keyboards may use diode matrices. TTL encoders. or MaS encoders. The
codes may be ASCII. EBCDIC. or a custom code. PROMs are often part of the encoding
CirCUitry.

The encoding Circuitry may do more than lust encode key

ROLLOVER
closures. It may also debounce the keys and handle "rollover." the
problem of more than one key being struck at the same time. Common ways of handling rollover are: "2-key rollover." whereby two keys (but not more) struck at the same
time are resolved Into separate closures. and "n-key rollover." whereby any number of
keys struck at the same time are resolved Into separate closures.

The encoded keyboard also prOVides a strobe With each data transfer. The strobe signals that a new closure has occurred. Figure 11-23 shows the interface between an encoded keyboard and the Z80 microprocessor. The rising edge of the strobe latches the
data Into the input port. We also tie the strobe to the B Side of the Pia so that the CPU
can determine when a rising edge has occurred. Of course, the B port of one Pia could
hold status Signals from up to eight ports. The software would then have to determine
which ports were active with a shifting and masking operation.
We have assumed In the program that the strobe signal is long enough for the CPU to
handle it in software. If it is not. the Signal will have to be latched and cleared (with
ROY) when the mput or output transfer occurs.
You may have to watch the polarity of the strobe. since the Pia always reacts to a rising
edge. An Inverter gate may be necessary.

11-69

Task: Input from keyboard.

Purpose: Walt for the rising edge of a strobe at the B port of a Pia and then place the
data from Port A Into the Accumulator.
Flowchart:

Read
status port

Read
status port

Read
data port

The hardware must hold the control lines in a logic one state during reset to prevent the
accidental setting of status flags.

Source Program:

SRCHL:

SRCHH:

LD
OUT
LD
OUT
LD
OUT
IN
BIT
JR
IN
BIT
JR
IN
HALT

A,01001111B
(PIOCRAl,A
A, 11 001111 B
(PIOCRBl.A
A.OFFH
(PIOCRBl,A
A,(PIODRB)
STB.A
NZ,SRCHL
A(PIODRB)
STB.A
Z,SRCHH
A(PIODRA)

:MAKE PORT A INPUT

:MAKE PORT B CONTROL
:ALL PORT B LINES INPUTS
:EXAMINE STATUS PORT
:HAS STROBE LINE GONE LOW?
:NO, WAIT UNTIL IT HAS
:EXAMINE STATUS PORT AGAIN
:RISING EDGE FOUND?
:NO, WAIT UNTIL ONE OCCURS
:YES, FETCH DATA

11-70

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A

3E
4F
03
PIOCRA
3E
CF
03
PIOCRB
3E
FF
03
PIOCRB
DB
PIOORB
CB
STB
20
FA
DB
PIOORB
CB
STB
28
FA
DB
PIOORA
76

Instruction
(Mnemonic!

SRCHL.

SRCHH:

LO

A.01 001111 B

OUT

(PIOCRA),A

LO

A,11001111B

OUT

(PIOCRBl,A

LO

A.OFFH

OUT

(PIOCRBl,A

IN

A.(PIOORB)

BIT

STB.A

JR

NZ.SRCHL

IN

A.(PIOORB)

BIT

STB.A

JR

Z.SRCHH

IN

A.(PIOORA}

HALT

If the CPU repeats this routine. It will not fetch another character until the next rising
edge occurs on the strobe line. A continuing high level on the strobe line will be ignored.
STB depends on which bit of Port B IS used. Figure 11-23 shows bit 4 being used. but
bits O. 6. and 7 are. as usual. the easiest to examine. Try rewriting the program to use
the more accessible bit positions.
The second bvte of the Bit instructions depends on the value of STB but is not equal to
that value. For example. the second byte IS 4F16 if STB = 1. 5716 if STB = 2. etc.

11-71

A Digital-to-Analog Converter
Purpose: Send data to an 8-blt digltal-to-analog converter. which has an active-low

latch enable.
Oigital-to-analog converters produce the continuous signals required by solenoids.
relays. actuators. and other electrical and mechanical output devices. Typical converters consist of sWitches and resistor ladders with the appropriate resistance values 7
The user must generally provide a reference voltage and some other digital and analog
circuitry. although complete Units are becoming available at low cost.
Figure 11-24 describes the 8-bit Signetlcs NE5018 O/A converter. which contains an
on-chip 8-blt parallel data input latch. A low level on the LE (Latch Enable) input gates
the input data Into the latches. where it remains after LE goes high.
Figure 11-25 illustrates the Interfacing of the device to a Z80 microprocessor. Here the
A side of the Pia is used to generate the Latch Enable signal. The ROY line from the Pia
could be used In the mode where It is tied to the STS line to form a pulse lasting one
clock cycle. However. one clock cycle may not be long enough. since the NE5018 reqUires a 400 ns pulse. Furthermore. the polanty IS the opposite of that needed by the
NE5018.
Note that the Pia latches the output data. The data therefore remains stable during and
after the conversion. The converter typically requires only a few microseconds to produce an analog output. Thus. the converter latch could be left enabled if the port were
not used for any other purpose.
In applications where eight bits of resolution are not enough. 10- to 16-blt converters
can be used. Additional port logic IS required to pass all the data bits: some converters
provide part of this logiC.
The Pia here serves both as a parallel data port and as a serial control port. Of course. if
Port A is used for control. it could actually handle up to eight bits.
Task: Output to converter.
Purpose: Send data from memory location 0040 to the converter.

Flowchart:

} Data = 10040)

Sand data
to converter

Pulse
Latch Enable

11-72

Digital

LE

DB7
MSB

DOO

DBS

DB4

DBJ

DB2

DBl

GND

DBO

LSB

Latches and
Switch Drivers
j

VCC

VREF
Out

Sum
Node

R
>-e--oVOut

OAC Current
Output
I

VREF

Amp
Camp

Adj

o Analog

DAC Switches

--

GND

R
VREF

In

Bipolar
Offset

f"
VCC

All R values equal 5k n and are thermally matched

Figure 11-24 Signetics NE5018 D/A Converter

Data Bus
from CPU

Analog

Pia

B7
BO

NE5018
D/A

Output

Converter

i:f

A4

Figure 11-25. Interface for an a-bit Digital-to-Analog Converter

Source Program:
LD
OUT
SUB
OUT
LD
OUT
LD
OUT
IN
RES
OUT
SET
OUT
HALT

A.11001111B
(PIOCRAl.A

:MAKE PORT A CONTROL

:ALL PORT A PINS OUTPUTS

(PIOCRAl.A
A.00001111 B
(PIOCRBl.A
A.(40H)
(PIODRB},A
A.(PIODRA)
4.A
(PIODRAl.A
4.A
(PIODRA).A

:MAKE PORT B OUTPUT

:GET DATA
:SEND DATA TO DAC
:GET OLD CONTROL DATA
:BRING LATCH ENABLE LOW
:BRING LATCH ENABLE HIGH

11-74

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A

3E
CF
03
PIOCRA
97
03
PIOCRA
3E
OF
03
PIOCRB
3A
40
00
03
PIODRB
DB
PIODRA
CB
A7
03
PIODRA
CB
E7
03
PIODRA
76

Instruction
(Mnemonic)
LD

A.11001111B

OUT

(PIOCRAl.A

SUB
OUT

A
(PIOCRAl.A

LD

A.00001111 B

OUT

(PIOCRBl.A

LD

A.(40H)

OUT

(PIODRB).A

IN

A.(PIODRAl

RES

4.A

OUT

(PIODRAl.A

SET

4.A

OUT

(PIODRAl.A

HALT

The particular bit that must be set and reset depends. of course. on how the Latch Enable IS connected to the control port. Bit is often convenient to use for control purposes
since. if that bit is orlginallv cleared. It can be set with an INC Instruction and reset with
a DEC instruction.

We could use the automatic brief strobe from BACK if the Latch Enable were actlvehigh (and if this strobe were long enough when BACK IS tied back to B STB). The program would then be:
LD
OUT
LD
OUT
HALT

A.00001111 B
(PIOCRB).A
A.(40H)
(PIODRB).A

:MAKE PORT B OUTPUT

:GET DATA
;SEND DATA TO DAC AND ENABLE LATCH

An Inverter gate could produce an active-low signal. Note how many fewer instructions
are necessary.

11-75

Analog-to-Dlgltal Converter
Purpose: Fetch data from an 8-bit analog-ta-digital converter that requires an Initiate
Conversion pulse to start the conversion process and has a Data Valid line to
indicate the completion of the process and the availability of valid data.
Analog-ta-digital converters handle the continuous signals produced by vanous types
of sensors and transducers. 8 The converter produces the digital input which the computer requires.
One form of analog-to-digital converter IS the successive approximation device. which
makes a direct 1-bit companson during each clock cycle. Such converters are fast but
have little nOise Immunitv. Dual slope integrating converters are another form of
analog-to-digital converter. These devices take longer but are more resistant to noise.
Other techniques. such as the Incremental charge balancing technique. are also used.
Analog-to-digital converters usuallv reqUire some external analog and digital circuitry.
although complete units are becoming available at low cost.
Figure 11-26 shows the 8-bit Teledyne Semiconductor 8703 AID converter. The device
contains a result latch and tristate data outputs. A pulse on the Initiate Conversion line
starts conversion of the analog input; after about two milliseconds the result will go to
the output latches. and the Data Valid output will Indicate this by switching first low
and then high. Data is read from the latches by applying '0' to the ENABLE input.
Figure 11-27 shows the interface for the Z80 processor and the 8703 converter. 9 Port B
is used to prOVide an Initiate Conversion pulse (active-high) of sufficient length. The
Data Valid signal is tied to A STB so that Data Valid going low and then high will latch
the converted data into Port A. The Data Valid signal is also tied to a bit of Port B so that
the CPU can determine ItS value. The important edge on the Data Valid line IS the lowta-hlgh edge. which indicates the completion of the conversion. As in the case of the
encoded keyboard. additional circuitry will be necessary if the pulse on Data Valid is too
short to be handled in software. Note that we are using Port B here for both status and
control.

11-76

0. 1I'!'

~.

+5V

19
INITIAT E
CONVERS ION

...r1-

VIN

21
1

.5V

-..I
-..I

lookfi
201<fi

loon

'

Zero

Adjust
5V

+p-

..!2P AMP

I"

'''''''''':~

~mparator

.......

16

-V

-.

Clock
and
Control
Logic

20pA

18

RREF'
~ 3201<0

Data

Counters

Output
Latches

;.

10

fo4-

8
9

11
12
LS8

la'8it

'r

22
23

ata

BUSY
Data Valid

GNO

Vss

13

5
6

Internal

lkO

-.1-

ENABLE

MS8

.))---------

- -2"

RIN 1Mfi' ::;::<1NT 68 pF

liN = 10pA
iA FoS.... 270pF

........

Binary
Output

VOO

,,17

RBiAS

20

lOOkO

fO'll'!'
l

VREF

l
-5V

.... '*
Components chosen for VIN (FS)

Figure 11-26. Teledyne 8703 AID Converter

lOV. VREF

-64V

Analog

Data Bus

to CPU

Input

<

PIO

AID

Ao
A STB

85

TBledvn. 8703

":7

82

Converter

----

Data

InitIate

Valid

Conversion

Figure 11-27. Interface for an 8-bit Analog-to-Digltal Converter

Task: Input from converter.
Purpose: Start the conversion process. wait for Data Valid to go low and then high.

and then read the data and store it in memory location 0040.
Flowchart:

Pulse lniti8te
Conversion

line

Read data from

data Input port

100401 = Data

Note that here the PIO serves as a parallel data port. a serial status port. and a serial
control port.

11-78

Source Program:
LD
OUT

LD
OUT
LD
OUT
LD
OUT
SUB
OUT
WTLOW: IN
BIT

JR
WTHI:

IN
BIT

JR
IN
LD
HALT

A.01001111 B
(PIOCRA).A
A.11 001111 B
(PIOCRBl.A
A.00001111 B
(PIOCRBl,A
A.00100000B
(PIODRB),A
A
(PIODRB).A
A. (PIODRB)

:MAKE PORT A INPUT

:MAKE PORT B CONTROL
:B4-7 OUTPUT. BO-3 INPUT
:SEND INITIATE CONVERSION HIGH
:SEND INITIATE CONVERSION LOW
:HAS DATA VALID GONE LOW?

2.A
NZ.WTLOW
A.(PIODRB)

:NO. WAIT
:IS DATA AVAILABLE?

2.A
Z.WTLOW
A.(PIODRA)
(40Hl.A

:NO. WAIT
:YES. FETCH DATA FROM CONVERTER
:SAVE CONVERTER DATA

11-79

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A
001B
001C
0010
001E
001F
0020
0021
0022
0023
0024

3E
4F
03
PIOCRA
3E
CF
03
PIOCRB
3E
OF
03
PIOCRB
3E
20
03
PIOORB

Instruction
(Mnemonic)
LO

A.01001111 B

OUT

(PIOCRAl.A

LO

A.11 001111 B

OUT

(PIOCRBl.A

LO

A.00001111 B

OUT

(PIOCRBl.A

LO

A.00100000B

OUT

(PIOORBl.A

97

SUB

03
PIOORB
DB
PIOORB
CB
57
20
FA
DB
PIOORB
CB
57
28
FA
DB
PIOORA
32
40
00
76

OUT

(PIOORBl.A

WTLOW: IN

WTHI:

A.(PIOORB)

BIT

2.A

JR

NZ.WTLOW

IN

A.(PIOORB)

BIT

2.A

JR

Z.wTHI

IN

A.(PIOORA)

LO

(40H).A

HALT

One approach to configUring PI Os IS to use the repeated Block Output Instruction OTIR
and a table In memory containing the words to be sent to the Control register. A typical
routine would be:
LO
LO
LO
OTIR

B,LENG
C,PIOCR
HL,CTLTAB

;COUNT NUMBER OF CONTROL WORDS

;GET CONTROL PORT NUMBER
:STARTING ADDRESS OF PIO CONTROL TABLE
:CONFIGURE PIO

In fact. another table (or the Stack) could be used to hold the number of control words
and the port number for each PIO.

11-80

One state
Zero state

_----...
l'

-...-----...
Start
Bit

'0'

'0'

'0'

'"

'0'

,.-----.,,~-...-~
Panty
Stop
Stop

7 Data Bits

Bit

Bit

Bit

Ol.racter is ASCII 'E' with odd parity (45 hex;.

Remember that the transmission order is Start bit
~L~Q~t~~~a~~~~~&~~
bit, Stop ~ ('
Stop bit ('1').

Figure 11-28. Teletypewriter Data Format

A Teletypewriter (TTY)
Purpose: Transfer data to and from a standard 10-character-persecond serial teletypewriter,

The common teletypewriter transfers data In an asynchronous

serial mode. The procedure is as follows:
1)
2)
3)
4)
5)

The line is normally In the one state,

A Start bit (zero bit) precedes each character,
The character is usually 7-blt ASCII with the least Significant
bit transmitted first.
The most Significant bit IS a Parity bit, which may be even,
odd. or fixed at zero or one.
Two stop bits (logic one) follow each character,

STANDARD
TTY

CHARACTER
FORMAT

Figure 11-28 shows the format. Note that each character requires the transmission of
eleven bits. of which onlY seven contain information, Since the data rate is ten characters per second, the bit rate IS 10 x 11, or 110 Baud, Each bit therefore has a width of
1/110 of a second, or 9.1 milliseconds, This width is an average: the teletypeWriter
does not maintain it to any high level of accuracy.
For a teletypewriter to communicate properly with a computer. the follOWing procedures are necessary.
Receive (flowcharted In Figure 11-29):
Step 1)
Step 2)
Step 3)

Step 4)
Step 5)

TTY

RECEIVE

Look for a Start bit (a logic zero) on the data line,

MODE
Center the reception by waiting one-half bit time. or 4.55
milliseconds,
Fetch the data bits. waiting one bit time before each one. Assemble the data
bits into a word by first shifting the bit to the Carry and then circularly shifting
the data with the Carry, Remember that the least Significant bit is received
first.
Generate the received Parity and check' it against the transmitted Parity, If
they do not match. Indicate a "Parity error,"
Fetch the Stop bits (waiting one bit time between Inputs). If they are not correct (if both Stop bits are not onel. Indicate a "framing error,"

11-81

Gat Input data

Panty
error

Wait

one-ha~

Count =2

bit tune

Count

Oats

Wait one bit time

Get input data

Wait one bit time

Get ,"put data

Carry = Input date
Shift dete nght
with Carry

Count

Framing

= Count-1

error

Count

= Count -

Generate
received panty

Figure 11-29. Flowchart for Receive Procedure

11-82

Task 1: Read data.

Purpose: Fetch data from a teletYpewriter through bit 7 of a PIO data port and place
the data into memory location 0060. For procedure. see Figure 11-29.
Source Program:
(Assume that the serial port is bit 7 of the PIO and that no panty or framing check
necessary)

WTSTB:

RCVB:

LD
OUT
IN
RLA

A.01001111B
(PIOCRA),A
A.(PIODRA)

JR

C.wTSTB
DHALF
D.10000000B
DFULL
A.(PIODRA)

CALL
LD
CALL
IN
RLA

RR
JR
LD
LD
HALT

o
NC.RCVB
A.D
(60H).A

:MAKE PORT A INPUT

:READ SERIAL LINE
:IS THERE A START BIT?
:NO. WAIT UNTIL THERE IS
:YES. DELAY HALF BIT TIME TO CENTER
:COUNT WITH BIT IN MSB
:WAIT 1 BIT TIME
:READ SERIAL LINE
:MOVE BIT TO CARRY
:MOVE BIT TO ASSEMBLED WORD
:CONTINUE IF COUNT BIT NOT IN CARRY

(Delay program)
DHALF'

ORG
PUSH
LD

JR
DFULL:
DLY16:
DLY1.

PUSH
LD
LD
DEC

JR

30H
DE
D.B
DLY16
DE
0.16
E.8DH

:SAVE OLD REGISTERS

;FULL BIT LENGTH COUNT
;DELAY 1/16TH BIT TIME

E
NZ.DLY1

DEC

JR

NZ.DLY16
DE

POP
RET

:SAVE OLD REGISTERS

:HALF BIT LENGTH COUNT

;RESTORE OLD REGISTERS

Remember that bit 0 of the data is received first.

11-83

IS

Object Program:

Memory Address
(Hex)
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A
001B
001C
0030
0031
0032
0033
0034
0035
0036
0037
0038
0039
003A
0038
003C
003D
003E
003F
0040
0041

Memory Contents
(Hexl
3E
4F
D3
PIOCRA
DB
PIODRA
17
38
FB
CD
30
00
16
80
CD
35
00
DB
PIODRA
17
CB
lA
30
F6
7A
32
60
00
76
D5
16
08
18
03
D5
16
10
1E
8D
1D
20
FD
15
20
F8
Dl
C9

Instruction
(Mnemonic)
A.Ol00ll11 B
LD
OUT
WTSTB: IN

(PIOCRA),A
A.(P1ODRA)

RLA
JR

C,WTSTB

CALL

DHALF

LD

D.10000000B

CALL

DFULL

IN

A.(PIODRAI

RLA
RR

JR

NC.RCVB

LD
LD

A.D
(60H).A

HALT
PUSH
LD

DE
D.8

JR

DLY16

DFULL:

PUSH
LD

DE
D.16

DLY16:

LD

E.8DH

DLY1:

DEC
JR

E
NZ.DLY1

DEC
JR

D
NZ.DLY16

POP
RET

DE

RCVB:

DHALF:

11-84

This program assumes that the Stack can be used for subroutine calls, i.e., the monitor
must initialize the Stack Pointer. Otherwise you will have to initialize the Stack Pointer
as shown in Chapter 10.
The constants for the delay routine were calculated Just as shown earlier in this chapter.
You might try determining them for yourself. The delays do not have to be highly accurate because the reception is centered, the messages are short, the bit rate is low, and
the teletypewriter is not highly accurate itself.
How would you extend this program to check for the two stop bits? They must both be
one or a framing error has occurred.
You can extend this program to check odd parity by replacing the LD A,D Instruction
with the sequence:
A

SUB
AND

JP

PE,PRERR

:IS PARITY ODD?

;NO, PARITY ERROR HAS OCCURRED

11-85

Carry = 0 (Start bitl

Get output data

Shift data left

circularly with Carry
Count =11

SeIl\j

data to

Output Port

Shift data nght

circularly with Carry

Carry

= 1 {Stop bitl

Wait 1 bit time

Count = Count - 1

Figure 11-30. Flowchart for Transmit Procedure

Task 2: Wnte data.

Purpose: Transmit data to a teletypewriter through bit 0 of a PIO data register. The
data is in memory location 0060.

TTY

Transmit (flowcharted in Figure 11-30)

Step 1)
Step 2)
Step 3)
Step 4)

Transmit a Start bit O.e.. a logic zero).

Transmit the seven data bits. starting with the least significant bit.
Generate and transmit the Panty bit.
Transmit two Stop bits (i.e.. logic ones).

TRANSMIT

MODE

The transmission routine must wait one bit time between each operation.

11-86

Source Program: (Assume that parity need not be generated)

: MAKE PIO INTO OUTPUT PORT
LD
OUT

A.00001111 B
(PIOCRB),A

:MAKE PORT B OUTPUT

: GET DATA AND CLEAR START BIT

LD
ADD
LD

A, (60H)

:GET DATA
:SHIFT LEFT AND FORM START BIT
;COUNT = 11 BITS

A.A
B.ll

. TRANSMIT A BIT AND UPDATE DATA

TBIT:

OUT
RRA
SCF

(PIODRB),A

:TRANSMIT A BIT
:UPDATE FOR NEXT BIT
:FORM STOP BIT (LOGIC ONE)

: DELAY 9.1 MS AND COUNT BITS

CALL
DJNZ
HALT

DFULL
TBIT

The DFULL subroutine

transferred first.

IS

:DELAY 9.1 MS
:COUNT DOWN 11 BITS

the same as before. Remember that bit 0 of the data must be

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
DODD
OOOE
OOOF
0010
0011
0012
0013

3E
OF
D3
PIOCRB
3A
60
00
87
06
DB
D3
PIODRB
IF
37
CD
35
00
10
F7
76

Instruction
(Mnemonic)

TBIT:

LD

A.OOOOllll B

OUT

(PIOCRB),A

LD

A, (60H)

ADD
LD

A.A
B.ll

OUT

(PIODRB),A

RRA
SCF
CALL

DFULL

DJNZ

TBIT

HALT

11-87

ADD A.A clears the least significant bit so that It can be used as the start bit. The most
significant bit is saved in the Carry. In actual applications. the startup routine should
place a logic '1' on the teletypewriter line after configuration since that line should normally be In the mark (one) state.
Each character consists of 11 bits. starting with a start bit (zero) and ending with two
stop bits (ones).
This program can easily be extended to generate 7-bit characters with odd parity In the
most significant bit. The parity generation routine (to be Inserted after LD A.(60H)) IS:

STBIT:

ANA
JP
SET
ADD

A
PO.STBIT
7.A
A.A

;IS PARITY ODD?

;YES. NO PROBLEM
;NO. MAKE IT ODD BY SETTING MSB
;SHIFT LEFT AND FORM START BIT

How would you generate even panty?

These procedures are sufficiently common and complex to merit a

IUART
special LSI device: the UART. or Universal Asynchronous
Recelver/Transmitter. 10 The UART will perform the reception procedure and provide
the data In parallel form and a Data Ready signal. It will also accept data in parallel
form. perform the transmission procedure. and provide a Penpheral Ready signal when
It can handle more data. UARTs may have many other features. Including:
1)
2)
3)

Ability to handle various bit lengths (usually 5 to 8l. panty options. and numbers of
Stop bits (usually 1. 1-1/2. and 2).
Indicators for framing errors. parity errors. and "overrun errors" (failure to read a
character before another one IS received).
RS-232 11 compatibility: I.e.. a Request-to-Send (RTS) output signal that indicates
the presence of data to commUnications equipment and a Clear-to-Send (CTS) Input signal that indicates. in response to RTS. the readiness of the communications
equipment. There may be provisions for other RS-232 signals. such as Received
Signal Quality. Data Set Ready. or Data T~rmlnal Ready.

4)

5)
6)

Tristate outputs and control compatibility with a microprocessor.

Clock options that allow the UART to sample incoming data several times In order
to detect false Start bits and other errors.
Interrupt facilities and controls.

UARTs act as four parallel ports: an input data port. an output data port. an input
status port. and an output control port. The status bits Include error Indicators as well
as Ready flags. The control bits select vanous options. UARTs are Inexpensive ($5 to
$50. depending on features) and easy to use.

11-88

THE

zeo SERIAL INPUT/OUTPUT DEVICE (SIO)

The ZSO Serial Input/Output Device or 510 (see Figure 11-31) IS a complete communications contrdller specifically designed for use in ZSO-based microcomputers. It
can serve a variety of communications functions. but we will only discuss ItS use as a
simple asynchronous recelverltransmitter. 12
The 510 has two complete channels (A and B) which can both
receive and transmit serial data (see Figure 11-32). Channels that
can receive and transmit simultaneously are called full-duplex.
Alternatives include half-duplex (able to transmit and receive, but not at the same
time). receive-only, and transmit-only.
An 510 occupies four Input port addresses and four output port
addresses. The B/A (Channel B or A Select) and C/O (Control or
Data Select) li.,es choose one of the four ports as described in Table 11-7. Most often, designers attach address bit AO to the B/A Input and address bit
A1 to the C/O input. The 510 then occupies four consecutive port addresses as described in the last column of Table 11-7
As with the PIO, SIOs have more control registers than adADDRESSING
dresses. In fact. each 510 has eight registers in each chan810 READ AND
WRITE REGISTER
nel for control and three registers for status. Figure 11-33
contains diagrams of each control or Write register: Figure
11-34 contains diagrams of each status or Read register. Two transfers are reqUired to
read or write any of the registers except Write Register O. The first transfer (written Into
Write Register 0) contains three bits that direct the next transfer to or from the selected
register. Note, In Figure 11-33, that these three bits occupy the three least Significant
bit positions and that zeros in the other bit posilions indicate a byte that has no function
other than addressing.

11-89

Senal Data
Internal

+5V GND

<l>

Channel Clock

Channel A

Control

SYNC

LogIc

WAITfRDY

+++
Data
Discrete

CPU

Modem or

Control

Bus

Other

and

I/O

Controls

Status

Control

Senal Data
Internal

....II---Channel Clock
Channel B

Control

Logic

jnterrupt Control

Lines

Figure 11-31. Block Diagram of the ZBO SIO

Table 11-7 SIO Addresses

CONTROL OR
DATA SELECT

CHANNEL B OR A
SELECT

REGISTER
ADDRESSED

PORT ADDRESS
(STARTING WITH SIOADDl

0
0

1
1

Data Register A
Data Register B
Control A
Control B

SIOADD
SIOADD+1
SIOADD+2
SIOADD+3

1
1

The port addresses assume that C/O

11-90

IS

tied to A1 and B/A to AO

TxD

TxC

II
XMIT

Shift and
Bit Insert

CRC

"-

"

SYNC

Generator

RegIsters
XMIT

Buffer

.:

i"

Channel

Internal Bus

Control

II

and

i"

Status

'< '7
REC

FIFO
CRC

SYNC

Checker

Detect

REC

Shift and
Bit Stnp

-v

I 1

RxD

RxC

Figure 11-32. Block Diagram of $10 Channel

11-91

Write Regllte..

The ZSO 510 contains eight registers in each channel that are programmed (written into) by the system software
to configure the functional personality of each channal. All Write registers. with the exception of Write Aegister 0,
require two bytes to be property programmed. The first byte contains three bits that pOint to the selected register
(DO~02); the second byte IS the actusl control word that is being written to that register to configure the SIO.

Write Register 0 is 8 special case. RESET (either internal command or external input) will initialize the 510 to Write
Aegister O. All basic commands (CM02-CMDO) and CAC controls (CACO, CACn cen be accessed with a single byte
using Write Register O.
Contained in the first byte of any Write register access are the basic commands (CM02-CMDO) and the CAC controls (CACO,CAC1) so that meximum system control and flexibility is maintained.

Write Regilt" 0

0
0
0
0

0
0

0
1

Aegister 1

Aegister 2

0
0

Aegister 3
Aegister 4

Aegister 5

RegIster 6
Register 7

1
1

o
o

Register 0

0
0

Null Code

Send Abort (SOLC)

Reset External or Status interrupts
Channel Reset

o
o

1
1
1
1

0
0
1
1

o
,

Aeset Ax Interrupt on First Charseter

Rsset Tx Interrupt Pending

Error Aem

Aetum from Interrupt (Ch-A Onlyl.

Null Code

1
0

Aeset Ax CAC Checker

Aeset Tx CAC Generstor

Aeset CAe/SYNCS Sent/Sending latch

Write Regllter 1

I 07 106 105 1D4 I 03 I 02 I 01 I DO I

~

External Interrupt Enable

Tx Interrupt Enable
Status Affects Vector (Ch-B Only;

0
0

Ax Interrupt Disable
Rx Interrupt on Rrst Charactsr Only or Error
Interrupt on All Rx Characters (Parity Affects Vector)
Interrupt on All Ax Charscters (Parity Doea Not Affect Vector)
WAIT/AEADY on AIT
WAIT FN/AEAOY FN
WAIT/AEAOY Enable

Figure 11-33. SIO C9ntrol or Write Registers

.11-92

Write Register 2

I 07 106 105 I 04 103 102 101 1DO I

~

Write Register 3

I 07 106 105 I 04 103 102 I 01

o
o

Rx 5 Bits/Character

Rx 6 Bits/Character
Rx 7 Bits/Charactar
Rx 6 Bits/Character

VO
V1
V2
V3
V4
V5
V6
V7

Interrupt
Vector

I DO I

Rx Enable
SYNC Character Load Inhibit
Address Search Moda (SOLC)
Rx CRC Enable
Enter Hunt Moda
Auto Enables

Write Register 4

L...

o
o
1
1
o

o
1
1

o
o

1
1

0
1
0
1

0
1
0
1

0
1
0
1

' - - - - - Parity Enable

Parity Even/Odd

SYNC Modss Enable

1 Stop Bit/Character
1-1/2 Stop Bits/Character
2 Stop Bits/Character

8 Bit SYNC Character

16 Bit SYNC Character
SOLC Mode (01111110 SYNC Ragl
External SYNC Mode

x 1 Clock Mode
x16 Clock Mode
x32 Clock Mode
x64 Clock Mods

Figure 11-33. 510 Control

or Write Registers (Continued)

11-93

Write Reglater 6

I 07 106 1OS 104 1 03 102 101 1DO I

Tx CRC Eneble

RTS
SOlC/CRC-16
Tx Eneble
Send BREAK
0
0
1
1

Tx
Tx
Tx
Tx

1
0
1

S Bits {or lessl/Cheracter

7 Bits/Character
6 Bits/Cherecter
6 Bits/Cheracter

OTR

Write Regleter 6

I 07 1D61

Dsl 041 D3 1 021 DIlDO

II

SYNC Bit 0
SYNC Bit 1
SYNC Bit 2
SYNC Bit 3

Aiso SOlC Address Fiei

SYNC Bit 4
SYNC BitS
SYNC Bit 6
SYNC Bit 7

Write Reglater 7

I 1 I I 1 I I IDO I
D7

D6

DS

D4

03
j

D2

Dl

SYNC Bit 6
SYNC Bit 9
SYNC Bit 10
SYNC Bit 11
SYNC Bit 12
SYNC Bit 13
SYNC Bit 14

For SOle it must be programmed

to "01111110" for F1eg Recognition

SYNC Bit lS

Figure 11-33. SIO Control or Write Registers (Continued)

11-94

Read Raglate..

The zao SID contains threa registers thet can be read to obtain the status of each channel. Status information includes error conditions, interrupt vector, and standard communication interface protocol signals. To read the con~
tents of e selected Raad registar. the system software must first write out to the SID the byte containing pointer
information (OO~D2) in exactly the same manner 8S s Write register operation. Then. by issuing a READ operation,
the contents of the addressed Read/Status register can be read by the

zao CPU.

The real power in this type of command structure is that the programmer has complete freedom, after pointing to
the selected register. of either reading or writing to initialize or test that register. By designing software to Initialize
the

zao Sig in a modular. structured fashion. the programmer can use the powerful ZSO BlOCk I/O instructions to

significantiv simplify and speed his software development and debug.

Read Reglater 0

I 07 106 105 104 103 I 02 101 1DO I

~

Rx Character Available
Interrupt Pending (Ch-A Oniyi
Tx Buffer Empty

OCO
SYNC/HUNT

CTS
Sending CRC/SYNCS
BREAK/ABORT

Read Register 1

I 07 106 105 104 103 102 101 1DO I

~
An Sent
I-Field Bits

in

1
0
0
1
1
0

0
1
0
1

0
0
0
1

1
1
1

PreVIO us

I-FiBld Bits In

Second Previous

Byte

Bvte

0
0
0
0
0
0

Ftesidue Data for

8 Rx Bits/Character

4
5
6

Programmed

6
6
8

Panty Error
Ax Overrun Error
CRC/Framlng Error

End of Frame (SOLCI

Special Rx
Condition

Interrupts

Figure 11-34. 510 Status or Read Registers

11-95

Reed Register 2 (Channel B Onlyl

I07/ 06/ osl 041 031 021 01 IDO I

VO
VI
V2
V3
V4

Interrupt
Vector

VS
V6

V7

Figure 11-34. SIO Status or Read Registers (Continued)

11-96

Note the following special features of the SIO:

1)

2)

3)

4)

Input and output Instructions address physically distinct

SPECIAL
registers. There is no way to read the control registers or write
FEATURES
into the status registers.
OFSIO
All control registers for a channel share a single port address.
Thus two bytes are required to change the contents of any control reg'ister except
Register O.
RESET Initializes the SIO to Write Register O. It also disables
both receivers arrd transmitters. deactivates all control sig~
nals. and disables all interrupts. We will discuss the SIO Interrupt system in Chapter 12.
The SIO must be configured before It can be used. The easiest way to do this is by
placing the required bytes into a table and using the repeated Block I/O instruction.
The table must Include both the bytes needed to address the various registers and
the data that must be placed Into them. A typical routine would be:

ISiOl

LD
LD
LD
OTIR
5)

6)

7)
8)

B.LENG
C.SIOCRA
HL.CTLTAB

;NUMBER OF WORDS IN TABLE

;PORT NUMBER
;START OF CONTROL TABLE
;CONFIGURE SIO

The RS-232 signals are all active-low. However. the SIO control bits for these signals are active-high O.e.. a logic '1' In a control bit sends an RS-232 signal lowl.
The SIO requires an external clock. In asynchronous communications at 110 Baud.
1760 Hz is usually supplied and the X 16 mode is used. The SIO will sample the bits
at the clock frequency for synchronization and to avoid false start bits caused by
nOise on the line.
The Data Ready (Rx Character Available) flag IS bit 0 of Read Register O. The Peripheral Ready (Ix Buffer Empty) flag IS bit 2 of Read Register O.
Error status bits (panty. overrun. and framing) are in Read Register 1.

11-97

EXAMPLES
Teletypewriter 1/0 via a USART
Task 1: Read from teletypewriter through SIO
Purpose: Receive data from a teletypewriter through an SIO and place the data into
memory location 0040. The data IS 7-bit ASCII with odd parity.
Source Program:

WAITD:

LD
OUT
LD
OUT
LD
OUT
LD
OUT
SUB
OUT
IN
RRA
JR
IN
LD
HALT

AA
(SIOCRAl.A
A.01 000001 B
(SIOCRA).A
A.3
(SIOCRA).A
A.Ol000001B
(SIOCRAl.A

A
(SIOCRA).A
A.(SIOCRA)
NC,WAITD
A.(SIODRA)
(40Hl.A

:ACCESS WRITE REGISTER 4

:X16 CLOCK MODE. ODD PARITY
:ACCESS WRITE REGISTER 3
:7 BIT CHARACTERS. ENABLE RECEIVER
:ACCESS READ REGISTER 0
;GET STATUS
:IS DATA AVAILABLE?
:NO. WAIT
:YES. GET DATA
:SAVE DATA IN MEMORY

11-98

Object Program:

Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A
001B
001C
0010

3E
04
03
SIOCRA
3E
41
03
SIOCRA
3E
03
03
SIOCRA
3E
41
03
SIOCRA
97
03
SIOCRA
DB
SIOCRA
1F
30
FB
DB
SIOORA
32
40
00
76

Instruction
(Mnemonic)

WAITO:

LO

AA

OUT

(SIOCRA),A

LO

A,01000001B

OUT

(SIOCRA),A

LO

A,3

OUT

(SIOCRA),A

LO

A,01 000001 B

OUT

(SIOCRA),A

SUB
OUT

A
(SIOCRA),A

IN

A,(SIOCRA)

RRA
JR

NC.WAITO

IN

A, (SIOORA)

LO

(40H),A

HALT

11-99

The program establishes Write Register 4 as follows:

Bits 7 and 6 = 01 to select X16 clock mode (1760 Hz
must be supplied)
Bit 1 = 0 to select odd parity
Bit 0 = 1 to enable parity checking

EXAMPLE

OFSIO
CONFIGURATION

The program establishes Write RegIster 3 as follows:

Bits 7 and 6 = 01 for 7 bIts per character
Bit 0 = 1 to enable the receiver
The received data status bit IS bit 0 of Read Register O.
Note that any errors found will be reported in Read Register 1:
Bit 6 = 1 for a framing error (no stop bit>
Bit 5 = 1 for an overrun error (more data received before
previous data read)
Bit 4 1 for a parity error

Try adding an error checking routine to the program. Set

(0061)

0
1
2
3

if no errors occurred
if a parity error occurred
if an overrun error occurred
if a framing error occurred.

Note that the receiver always checks for one stop bit.

11100

SIO
ERROR

STATUS

Task 2: Wnte to teletype through SIO.

Purpose: Send data from memory location 0040 to a teletypewriter through an SIO.
The data is 7-blt ASCII with odd panty.
Source Program:

WAITR:

LD
OUT
LD
OUT
LD
OUT
LD
OUT
SUB
OUT
IN
BIT

JR
LD
OUT
HALT

AA
(SIOCRA},A
A.01001101B
(SIOCRA},A

:ACCESS WRITE REGISTER 4

:X16 CLOCK MODE. 2 STOP BITS. ODD PARITY

A.5

:ACCESS WRITE REGISTER 5

(SIOCRA},A
A.00101000B
(SIOCRA},A

:7 BIT CHARACTERS. ENABLE TRANSMITIER

:ACCESS READ REGISTER 0

(SIOCRA},A
A. (SIOCRA)

2.A
Z.WAITR
A. (40H)
(SIODRA},A

:GET STATUS
:IS TRANSMITIER READY?
:NO. WAIT
:YES. GET DATA
:AND TRANSMIT IT

11-101

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A
001B
001C
OOlD
001E

3E
04
03
SIOCRA
3E
40
03
SIOCRA
3E
05
03
SIOCRA
3E
28
03
SIOCRA
97
03
SIOCRA
OB
SIOCRA
CB
57
28
FA
3A
40
00
03
SIOORA
76

Instruction
(Mnemonic)

WAITR:

LO

A,4

OUT

(SIOCRA).A

LD

A,01001101B

OUT

(SIOCRA).A

LO

A,5

OUT

(SIOCRA).A

LD

A,00101000B

OUT

(SIOCRA),A

SUB
OUT

A
(SIOCRA).A

IN

A,(SIOCRA)

BIT

2,A

JR

Z,WAITR

LO

A,(40H)

OUT

(SIOORA).A

HALT

The program establishes Write Register 4 as follows:

Bits 7 and 6 = 01 to select X16 clock mode (1760 Hz must be supplied)
Bits 3 and 2 = 11 to add 2 stop bits to each character
Bit 1 = 0 to select odd parity
Bit 0 = 1 to enable parity generation
The program establishes Write Register 5 as follows:
Bits 6 and 5 = 01 for 7 bits per character
Bit 3 = 1 to enable the transmitter
The transmitter status bit is bit 2 of Read Register 1.

11102

STANDARD INTERFACES
Other standard Interfaces besides the TTY current-loop and
RS-232 can also be used to connect peripherals to the microcomputer. Popular ones include:
1)

The senal RS449. RS422, and RS423 Interfaces. 13

2)

The 8-bit parallel General Purpose Interface Bus, also known as IEEE-488 or
Hewlett-Packard Interface Bus (HPIBl.14
The S-l 00 or Altalr/lmsai hobbYist bus. 15 This is also an 8-bit bus.
The Intel Multibus. 16 This IS another 8-blt bus that can, however, be expanded to
handle 16 bits In parallel.

3)
4)

PROBLEMS
6) Separating Closures from an Unencoded Keyboard
Purpose: The program should read entnes from an unencoded 3 x 3 keYboard and
place them Into an array. The number of entnes required IS In memory location 0040 and the array starts In memory location 0041.

Separate one closure from the next by waiting for the current closure to end. Remember
to debounce the keyboard (this can be Simply alms wait).
Sample Problem:

(0040)

04

Entnes are 7. 2. 2, 4
Resu It: (0041)
07
(0042)
02
(0043)
02
04
(0044)

7)

Read a Sentence from an Encoded Keyboard

Purpose: The program should read entnes from an ASCII keyboard (7 bits with a zero
Parity bit) and place them Into an array until It receives an ASCII penod (hex
2El. The array starts in memory location 0040. Each entry IS marked by a
strobe as In the example given under An Encoded Keyboard.
Sample Problem:

Result:

Entnes are H. E.
(0040)
48
(0041)
45
(0042)
4C
4C
(0043)
(0044)
4F
(0045)
2E

L. L. O..
H
E
L
L
0

11-103

8)

A Variable Amplitude Square Wave Generator

Purpose: The program should generate a square wave, as shown In the next figure,
using a D/A converter. Memory location 0040 contains the scaled amplitude
of the wave, memory location 0041 the length of a half cycle in milliseconds,
and memory location 0042 the number of cycles,

Assume that a digital output of 8016 to the converter results In an analog output of zero
volts, In general. a digital output of D results in an analog output of VOUT = -VREF
(D-80)/80 volts.
Sample Problem:

(0040)
(0041)
(0042)

AO (hex)
04
03

Result:
+VREF I

i
~:: -V:E~ ~----F----t- ---F----\- ----1-- ---\-----1iVREF

-4H'

-VREF

4ms

TI~

The base voltage IS 8016 = 0 volts.

Full scale IS 10016 = -VREF volts.
So A016 = (AO-80)/80 X -VREF = -VREF/4
The program produces 3 pulses of amplitude VREF/4 with a half cycle length of 4 ms,

9)

Averaging Analog Readings

Purpose: The program should take four readings from an AID converter ten milliseconds apart and place the average In memory location 0040. Assume that
the AID conversion time can be Ignored.
Sample Problem:

Readings are (hex) 86, 89, 81. 84

Result:

10)

(0040) = 85

A 30 Character-per-Second Terminal

Purpose: Modify the transmit and receive routines of the example given under A
Teletypewriter to handle a 30 cps terminal that transfers ASCII data with one
stop bit and even parity. How could you write the routines to handle either
terminal depending on a flag bit in memory location 0060: e.g., (0060) = 0
for the 30 cps terminal. (0060) = 1 for the 10 cps terminal?

11-104

REFERENCES
1.

Barnes. J., and V Gregory. "Use Microcomputers to Enhance Performance with

NOIsy Data." EON. August 20. 1976. pp. 71-72.

2.

Swanson. R.. "Understanding Cyclic Redundancy Codes." Computer Design.

November 1975. pp. 93-99: and McNamara. J. E.. Technical Aspects Qfi5iila
Communication. Digital Equipment Corp.. Maynard. Mass. 1977

3.

For example. the Z80 Direct Memory Access Controller (or DMA) for Z80 based
microcomputers IS described in An IntroductIOn to Microcomputers: Volume 2 Some Real Microprocessors.

4.

The TTL Data Book for Design Engineers. Texas Instruments Inc.. P O. Box 5012.
Dallas. Texas 75222. 1976.

5.

Dilatush. E.. "Special Report: Numeric and Alphanumeric Displays." EON. February 5. 1978. pp. 26-35.

6.

See Reference 4.

Hnatek. E. R.. A User's Handbook of DIA and AID Converters. Wiley. New York.
1976.

8.

See Reference 7

9.

See also D. Guzeman. "Marry Your p.P to Monolithic AIDs." Electronic Design.
-January 18. 1977. pp. 82-86.

10.

For a discussion of UARTs. see P Rony et al.. "The Bugbook Ila." E and L Instruments Inc.. 61 First Street. Derby. CT. 06418: or D. G. Larsen et al..
"INWAS: InterfaCing With Asynchronous Serial Mode," IEEE Transactions on Industrial Electronics and Control Instrumentation. February 1977, pp. 2-12. Also
see McNamara. Reference 2.

11.

The offiCial RS-232 standard IS available as "Interface between Data Terminal

Equipment and Data Communications Equipment EmplOYing Serial Binary Data
Interchange." EIA RS-232C August. 1969. You can find Introductory descriptions
of RS-232 In G. Pickles. "Who's Afraid of RS-232?," Kilobaud. May 1977. pp.
50-54 and in C. A. Ogdin. "Microcomputer Buses - Part II." Mini-Micro Systems.
July 1978. pp. 76-80. Ogdin also describes the newer RS-449 standard.

12.

The SIO is discussed more completely In Volume 3 of An Introduction to

Microcomputers; the follOWing reference describes ItS use as a data link controller: Welssberger. A. J.. "Data-Link Control Chips; Bringing Order to New Protocols," Electronics. June 8. 1978. pp. 104-112.

13.

Electronic Industries ASSOCiation. "Electrical Characteristics of Balanced Voltage

Digital Interface CirCUits," EIA RS-422. April 1975.
Electronic Industries ASSOCiation. "Electrical Characteristics of Unbalanced
Voltage Digital Interface CirCUits," EIA RS-423. April 1975.
Electronic Industries ASSOCiation. "General Purpose 37-Posltlon and 9-Posltlon Interface for Data Terminal EqUipment and Data Circuit Terminating EqUipment
EmplOYing Serial Binary Data Interchange," EIA RS-449. November 1977
Morris. D.. "ReVised Data Interface Standards." Electronic Design. September 1.
1977. pp. 138-141.
--

11-105

14.

Institute of Electrical and Electronics Engineers, "IEEE Standard Digital Interface

for Programmable Instrumentation," IEEE Std 488-1975-ANSI Me 1.1-1975.

J. B. Peatman, Microcomputer-Based Design, McGraw-HilI. New York,

1977; Loughry, D. C. and M. S. Allen, "IEEE Standard 488 and Microprocessor
Synergism," Proceedings olthe IEEE, February 1978, pp. 162-172.
15.

Morrow. G.. and H. Fullmer, "Proposed Standard for the S-100 Bus," Computer,
May 1978. pp. 84-89.

16.

Rolander, T.. "Intel Multibus Interfacing:' Intel Application Note AP-28. Intel CorporatIOn, Santa Clara. CA.. 1977

Smith. M. L.. "Build Your Own Interface:' Kilobaud, June 1977, pp. 22-28.

11-106

Chapter 12
INTERRUPTS
Interrupts are inputs that the CPU examines as part of each instruction cycle. These inputs allow the CPU to react to asYnchronous events in a more efficient manner than
polling each device. When Interrupts are utilized to initiate I/O. generally more hardware than ordinary. programmed I/O IS required. but this provides a faster and more
direct response. 1
Why use Interrupts? Interrupts allow events such as alarms. power
REASONING
failure. the passage of a certain amount of time. and peripherals
BEHIND
haVing data or being ready to accept data to get the immediate atINTERRUPTS
tention of the CPU. The programmer does not need to poll every
deVice. nor need the programmer worry about the system completely missing events.
An interrupt system IS like the bell on a telephone - I t nngs when a calliS received so
that you don't have to pick up the receiver occaSionally to see if someone IS on the line.
The CPU can go about ItS normal business (and get a lot more done!. When something
happens. the interrupt rouses the CPU and forces it to service the input before resuming
normal operations. Of course. this simple description becomes more complicated (just
like a telephone SWitchboard) when there are many Interrupts of varying importance
and there are tasks that cannot be interrupted.
The implementation of Interrupt systems vanes greatly.
Among the questions that must be answered to characterize a particular system are:

CHARACTERISTICS
OF INTERRUPT
SYSTEMS

1)

How many Interrupt Inputs are there?

2)
3)

How does the CPU respond to an interrupt?

How does the CPU determine the source of an Interrupt if the number of sources
exceeds the number of Inputs?
Can the CPU differentiate between Important and unimportant interrupts?
How and when IS the Interrupt svstem enabled and disabled?

4)
5)

There are many different answers to these questions. The aim of all the Implementations. however. is to have the CPU respond rapidly to interrupts and resume normal activity afterwards.
The number of Interrupt Inputs on the CPU chip determines the number of different
responses that the CPU can produce Without any additional hardware or software. Each
Input can produce a different Internal response. Unfortunately. most microprocessors
have a very small number (one or two. tYPically) of separate Interrupt Inputs.
The ultimate response of the CPU to an interrupt must be to transfer control to the correct Interrupt service routine and to save the current value of the Program Counter. The
CPU must therefore execute a Jump-to-Subroutine or Call instruction with the beginning of the Interrupt service routine as ItS address. ThiS action will save the return address In the Stack and transfer control to the interrupt service routine. The amount of
external hardware required to produce thiS response vanes greatly. Some CPUs internally generate the instruction and the address: others require external hardware to
form them. The CPU can only generate a different instruction or address for each separate input.
12-1

If the number of Interrupting devices exceeds the number of Inputs. the CPU will need extra hardware or software to identify the
source of the Interrupt. In the simplest case. the software can be a
polling routine which checks the status of the devices that may be
interrupting. The only advantage of such a system over normal polling IS that the CPU
knows that at least one device is active. The alternative solution IS for additional hardware to provide a unique data input (or "vector") for each source. The two alternatives
can be mixed: the vectors can Identify groups of inputs from which the CPU can identify a particular one by polling.

An interrupt system that can differentiate between important and

PRIORITY
unimportant interrupts IS called a "priority Interrupt system." Internal hardware can provide as many priority levels as there are inputs. External hardwarll can provide additional levels through the use of a Priority register and comparator.
The external hardware does not allow the Interrupt to reach the CPU unless its priority
IS higher than the contents of the Priority register. A PriOrity interrupt system may need
a special way to handle low-priority Interrupts that may be Ignored for long periods of
time.

Most interrupt systems can be enabled or disabled. In fact. most

ENABLING
CPUs automatically disable Interrupts when a RESET is performed
AND
(so that the programmer can configure the Interrupt system) and
DISABLING
on accepting an interrupt (so that the Interrupt will not interrupt
INTERRUPTS
its own service routine). The programmer may wish to disable interrupts while preparing or processing data. performing a timing loop. or executing a
mu Itl-byte operation.
An Interrupt that cannot be disabled (sometimes called a "nonmaskable InterrupC) may be useful to warn of power failure. an
event that obviously must take precedence over all other ac-

NON-MASKABLE
INTERRUPT

tiVitieS.

The advantages of interrupts are obvious. but there are also

disadvantages. These include:

DISADVANTAGES
OF INTERRUPTS

1)

Interrupt systems may require a large amount of extra

hardware.

2)

Interrupts still require data transfers under program control through the CPU. There
is no speed advantage as there is With DMA.

3)

Interrupts are random inputs. which makes debugging and testing difficult. Errors
may occur sporadically. and therefore may be very hard to find. 2
Interrupts may involve a large amount of overhead if many registers must be saved
and the source must be determined by polling.

4)

zso

INTERRUPT SYSTEM

The zao's internal response to an interrupt IS fairly complex. since there are three
different operating modes. The interrupt system consists of:
1)
2)

An active-low maskable interrupt Input (lNT) and an actlvelow non-maskable Interrupt input (NMi).
Two enable flipcflops (lFF1 and IFF2). IFF1 can be set or reset
to enable or disable interrupts. IFF2 serves as temporary
storage for IFF1 during non-maskable interrupts.

12-2

ZSO
INTERRUPT
INPUTS

The ZSO checks the current status of the interrupt system at the
end of each instruction cycle. If an interrupt is active and enabled.
the response is as follows: 3
1)

2)

lSO
INTERRUPT
RESPONSE

The CPU disables the Interrupt system by clearing IFF1. IFF2.

however. is left in its original state if a non-maskable interrupt has occurred. Note
that RESET clears both Interrupt flip-flops so that the system can be configured
before interrupts are enabled.
The CPU executes a special Interrupt Acknowledge cycle. distinguished by the M1
signal (operation code fetch) being active. MREO (memory request) inactive (so the
CPU will not perform ItS normal memory access). and 10RO (input/output request)
active so that an interrupt response vector can be placed on the Data Sus.

The remainder of the response depends on the Interrupt mode and the source.
Note in particular that the lSO will check for interrupts after each transfer or comparison in a Block Move. Block Compare. or Repeated Block I/O instruction.

The ZSO has the following special instructions for use with the
Interrupt system:
1)
2)
3)

4)

5)

6)

7)
S)

lSO INTERRUPT
INSTRUCTION

EI (Enable Interrupts) enables the maskable Interrupt by

setting the Interrupt flip-flops.
01 (Disable Interrupts) disables the maskable Interrupt by clearing the Interrupt flipflops.
RST (Restartlls a one-word Call instruction that saves the current value of the Program Counter in the Stack and Jumps to the address specified In the instruction.
Table 12-1 contains the various Restart instructions and their destination addresses. RST is often used in interrupt systems because it IS a one-word Instruction
that is easy to form and place on the Data Bus.
RETI (Return from Interrupt) acts exactly like a normal Return (RET) Instruction except that ZSO peripheral chips (PI Os. SIOs. and CTCs) recognize this Instruction and
use it as a notification that the current Interrupt service routine has been completed.
RETN (Return from Non-Maskable Interrupt) acts exactly like a normal Return (RET)
Instruction except that it loads IFF1 from IFF2 so as to restore the original state of
the interrupt system.
LD A.I loads the Accumulator with the contents of the I (Interrupt Vector) register.
This instruction (and LD A.R) also places IFF2 into the PIO bit of the Flag register.
That flag can then be tested or saved in the Stack.
LD LA loads the I (Interrupt Vector) register with the contents of the Accumulator.
1M (Set Interrupt Model determines the mode in which interrupts are serviced. The
three options are O. 1. or 2; these are described later in this chapter.

Non-Maskable Interrupt
The non-maskable interrupt is an edge-sensitive (negative
lSO
NON-MASKABLE
edge triggered) input. The processor therefore reacts only to
the edge of a pulse on thiS line. and the pulse will not interrupt
INTERRUPT
its own service routine. Non-maskable interrupts are useful for
applications that must respond to loss of power li.e.. must save data in a low-power
memory or switch to a backup battery). Typical applications are commUnications equipment that must retain codes and partial messages and test equipment that must keep
track of partially completed tests.

12-3

Table 12-1 The Restart (RST) Instructions

RST Instruction

Operation Code

(Mnemonic)

Hex!

(Hex)

Destination Address

(Decimal)

RSTO
RST 8
RST 10H
RST 18H
RST 20H
RST 28H
RST 30H
RST 38H

C7
CF

0000
0008

0
08

D7
OF
E7
EF
F7
FF

0010
0018
0020
0028
0030
0038

16
24
32
40
48
56

The Z80 responds to a non-maskable interrupt as follows:

1)

It clears IFF1, thus disabling all interrupts (but saving the old state of IFF1 In IFF2L

2)

It Ignores the next Instruction fetched from memory and instead jumps to memory
location 006616. saving the old value of the Program Counter in the Stack,

Remember that a RETN instruction at the end of the service routine will restore the old
state of IFF1 from IFF2,
We will not discuss the non-maskable Interrupt further. Henceforth. we will assume
that all Interrupt inputs are tied to INT

ZSO Interrupt Modes

The Z80 has three interrupt modes, The programmer can choose
any of these modes with the appropriate 1M instruction, On reset.
the processor always enters Mode O. The modes are:

Mode 0
In this mode. the CPU uses the data input during the Interrupt Acknowledge cycle as an
instruction, This mode is the same as the 8080 Interrupt response mode,4
The normal data input that must be proVided externally is a RST instruction (see Table
12-lL
RST is useful In Interrupt systems for the following reasons:
1)

It IS a one-word Instruction and so requires only one fetch

cycle,

RESTART
INSTRUCTION

2)

It prOVides eight different destination addresses or vectors,

3)

Its vectors are far enough apart to allow Jump instructions to reach the actual service routines,

4)

It is easy to form. since five of the bits are always '1.' An 8-to-3 encoder can proVide
the other three bits qUite easily,

RST has the following disadvantages:

1)

It cannot provide more than eight vectors,

2)

Its vectors are not far enough apart to allow space for entire Interrupt serYlce
routines,
Its vectors are In a fixed area of memory,

3)
4)

RST 0 has the same destination address as the RESET input and is therefore very
difficult to use, The system needs hardware to differentiate between RESET and
RST O. since the two cannot be distinguished by software alone,

12-4

Remember that RST saves the old Program Counter in the Stack Just as CALL does.

Mode 1
In this mode. the CPU Ignores the data input dUring the Interrupt Acknowledge cycle
and always executes RST 38H. thus Jumping to memory location 003816 and saving
the old Program Counter In the Stack. This mode IS equivalent to Mode 0 if the data input IS always RST 38H (FF16).
The advantage of this mode IS that no external hardware IS required. Its disadvantages
are that there IS no way to directly differentiate among Interrupt sources and the
destination address IS fixed. Mode 1 is useful In applications that have only one or two
interrupt sources and In which minimum hardware cost IS essential.

Mode 2
In this mode. the CPU uses the data input as part of an address from which to get the
starting address of the Interrupt service routine. When an interrupt IS accepted. the
CPU:
1)

Disables further Interrupts by clearing IFF1 and IFF2.

2)
3)

Stores the old Program Counter In the RAM Stack.

Forms a pOinter from the contents of Register I (eight MSBs) and the Data Bus Input
dUring the Interrupt Acknowledge cycle (eight LSBs). The least significant bit of
this pOinter IS forced to zero.

4)

Fetches an address from the two memory locations starting with the one referred to
by the pOinter (see Figure 12-1).

5)

Transfers control to the address obtained from memory.

Interrupt response in thiS mode requires 19 clock cycles.

The advantage of thiS mode is that It can provide a full page of 128 Interrupt service
vectors located anywhere In memory. The disadvantages of thiS approach are that the
interrupt response is slower and the system must be initialized. as follows:
1)
2)

The table of vectors must be loaded Into memorv if it IS not In ROM.

The I register must be loaded with the eight most significant bits (or page number)
of the table address. Note that RESET clears Register I. You can load I with a value
as follows:
LD
LD

3)

A.IPGNO
LA

;GET INTERRUPT PAGE NUMBER

:STORE IN VECTOR REGISTER

Interrupt Mode 2 must be set with the instruction 1M 2.

Mode 2 IS designed to work with Z80 PIOs. SIOs. and CTCs. PIO and SIO Interrupts are
described later In this chapter.

ZSO/SOSO INTERRUPT COMPATIBILITY

Mode 0 for the Z80 interrupt system IS. as mentioned. identical to the 8080 Interrupt
response. The 8080 does not have Interrupt Modes 1 or 2. although Mode 1 IS really Just
a special case of Mode O. The 8080 also has no NMI input.
The 8085 has additional Interrupt inputs. not available on either the 8080 or the Z80.
The 8085 also has a non-maskable interrupt (called TRAP) that forces a call to a
different address (2416) than that used by the Z80 NMI input.

12-5

,\
desired starting address
pointed to by:
Interrupt

8 Bits from

I Register

lo~

7 bits from

Peripheral

Service
Routine

low-order

high-order

> Starting
Address

Table

Figure 12-1. Forming an Interrupt Vector In Interrupt Mode 2

PIO INTERRUPTS
Most Z80 Interrupt systems involve PIOs. Each port of the PIO has
the follOWing features for use with Interruots:

PIO
INTERRUPTS

1)

An 8-blt Interrupt Vector register used to hold the eight least

significant bits of the table address formed by the CPU In Interrupt Mode 2.

2)

An Interrupt enable bit.

3)

An Interrupt Control register used to determine the logical operation performed and
the active polarity monitored for generating Interrupts In the control mode.

4)

An Interrupt Mask register used to determine which data lines will be monitored to
generate Interrupts in the control mode.

The Interrupt Vector register In each port can be accessed by WritIng a control word with a zero In ItS least Significant bit. as shown
below (see also Table 11-2):
07 06

05

04

03

02

01

PIO
INTERRUPT
VECTOR

00

Signifies this control word is an

Interrupt vector

A tYPical sequence to establish the value In this register is:

LD
OUT

A.IVECT
(PIOCRI.A

where IVECT has a '0' In ItS least Significant bit. The starting address for the Interrupt
service routine IS at address IVECT on the page aSSigned to the table of starting addresses for service routines.

12-6

07

06

04

05

03

02

......_ ..- _..v -

01

00

used In Mode 3 only

signifies Interrupt control word

Figure 12-2. Format for a PIO Interrupt Control Word

07

06

05

Only those port lines whose mask bit

04

IS

03

02

01

00

zero will be monitored for generatmg an mterrupt.

Figure 12-3. Format for a PIO Interrupt Mask

We can set the Interrupt control word in each port by writing a
control word with the format shown In Figure 12-2. If the port IS In
Mode 3. bits 06. 05. and 04 have the following meanings:
1)

06 = 1 means that all monitored I/O lines must become active

to cause an Interrupt (j.e.. a 10gicaiANOl. while 06 = 0 means

PIO
INTERRUPT
CONTROL
MODE

that any monitored I/O line becoming active will cause an Interrupt (i.e.. a logical
OR)
Note that an Interrupt occurs only if the logical equation IS true when Interrupts are
enabled or if It changes from false to true while Interrupts are enabled.
2)

31

05 defines the active polaflty (high or lowl of the monitored I/O lines 05
means active high. 05 = 0 means active low.
04 = 1 means that the next control word IS an Interrupt mask (Figure 12-3) Only
lines with a mask bit of zero will be monitored. 04 = 0 means that the mask does
not follow.

ENABLING AND
Bit 7 of the Interrupt control word determines the value of the
DISABLING PIO
Interrupt enable flip-flop for the port. Interrupts may be generINTERRUPTS
ated if the flip-flop IS set. Power-on resets thiS flip-flop. but
remember that the PIO has no RESET input. The Interrupt enable flip-flop may be set or reset without affecting the rest of the Interrupt control word
by writing a control word with the flip-flop value in bit 7 and 0011 in the four least significant bits.
Setting bit 4 of the Interrupt control word clears any pending Interrupts. ThiS can be
used to clear interrupts that may have occurred Inadvertently dUfing a reset.

12-7

Examples
1)

Interrupting output port with vector located at address

8016. Remember that the page number IS in the CPU I
register.
LD
OUT
LD
OUT
LD
OUT

A.OOOOllll B
(PIOCRB).A
A.80H
(PIOCRA).A
A, 10000011 B
(PIOCRBl.A

EXAMPLES OF
PIO INTERRUPT
CONFIGURATION

:MAKE PORT B OUTPUT

:VECTOR ADDRESS = 80 HEX
:ENABLE Pia INTERRUPT

An alternative that clears pending Interrupts as well as enabling interrupts from the
port IS:
LD
OUT

A, 10010111 B
(PIOCRA).A

:ENABLE Pia INTERRUPT

An interrupt will occur on the rising edge of STB.

2)

Interrupting Input port with vector located at address 6016.

LD
OUT
LD
OUT
LD
OUT

A,Ol 001111 B
(PIOCRA).A
A,60H
(PIOCRA).A
A,l 0000011 B
(PIOCRA).A

:MAKE PORT A INPUT

:VECTOR ADDRESS = 60 HEX
:ENABLE Pia INTERRUPT

An interrupt will occur on the rising edge of STB.

3)

Interrupting control port with vector located at address 4816. An Interrupt will,_\Je
generated if data lines A4 and A7 both go low.
LD
OUT
LD
OUT
LD
OUT
LD
OUT
LD
OUT

A.ll0011118
(PIOCRA).A
Al0001000B
(PIOCRA).A
AA8H
(PIOCRA).A
A.ll010111B
(PIOCRA).A
A,0111 0111 B
(PIOCRA),A

:MAKE PORT A CONTROL

:L1NES 4,7 INPUTS - OTHERS OUTPUTS
:VECTOR ADDRESS = 48 HEX
:ENABLE Pia INTERRUPT
:MONITOR LINES 4.7 ONLY

The Interrupt control word has:

bit 7
bit 6
bit 5
bit 4

1 to enable the Interrupt

1 to generate an Interrupt only if all monitored lines are or
become active (a logical AND)
a to specify that a logic '0' is the active state to be monitored
1 to Indicate that a mask word follows (and to reset pending
interrupts)

12-8

4)

Interrupting control port with vector located at address 2816. An Interrupt will be
generated if any of the data lines go high.
LD
OUT
LD
OUT
LD
OUT
LD
OUT
SUB
OUT

A.11001111B
(PIOCRB},A
A.OFFH
(PIOCRB},A
A.28H
(PIOCRB},A
A.10110111B
(PIOCRB},A
A
(PIOCRB},A

:MAKE PORT B CONTROL

:ALL LINES INPUTS
:VECTOR ADDRESS = 28 HEX
:ENABLE INTERRUPTS
:MONITOR ALL LINES

The Interrupt control word has:

bit 7
bit 6
bit 5
bit 4

1 to enable the Interrupt

0 to generate an interrupt if any monitored lines become active
(a logical OR)
1 to specify that a logic '1' IS the active state to be monitored
1 to indicate that a mask word follows (and to reset pending
interrupts).

Obviously a repeated Block Output instruction could be used to shorten these programs
considerably.
Each PIO also has a single interrupt output and enable signals
for daisy chaining. The INT output is active-low when the PIO
has an interrupt request. The enable signals are:
lEI" (Interrupt Enable In) - high if no other deVices of higher
priority are being serviced by a CPU interrupt service routine.
lEO (Interrupt Enable Out) from this PIO

high if lEI IS high and the CPU IS not servicing an interrupt

lEI and lEO can be used to form a daiSY chain (see Volume 1 of.ill
Introduction to Microcoml;1uters) in which PIOs and other devices
that are connected to the chain closer to the CPU can block interrupt requests from devices further from the CPU. The advantages
of the daiSY chain are:
1)

It Identifies each source uniquely.

2}
3)

It requires no other hardware.

It is easy to expand or rearrange in hardware.

The disadvantages of the daiSY chain are:

1)
2)
3)

DAISY
CHAINING
Pia
INTERRUPTS

Pia DAISY
CHAIN
SIGNALS

ADVANTAGES
AND
DISADVANTAGES
OF DAISY CHAIN
INTERRUPTS

It can be vaned or changed only in hardware.

It does not provide for eventual servicing of low pnonty interrupts.
It requires extra time because signals must ripple through the chain.

The Z80 automatically waits long enough for the signals to ripple through a chain of up
to four devices when operating in Interrupt Mode 2. Additional hardware can be added
to allow longer chains.

12-9

Note that a partlcu lar device in the chain operates as follows:

DEVICE
OPERATION

1)

It places Its Interrupt vector on the bus during an Interrupt

IN A DAISY
Acknowledge cycle only if it has a pending interrupt reCHAIN
quest and Interrupt Enable In is high (indicating no higher
priority devices are being serviced). Interrupt Enable Out IS
also set low. Within a device. Port A interrupts take precedence over Port B interrupts.

2)

It subsequently brings ItS Interrupt Enable Out high (enabling lower priority
devices) only if a RETI Instruction is executed while ItS Interrupt Enable In IS high.

Thus. a particular device will be serViced only when it has the highest priority request
and will block lower-priority requests until ItS service routine has been completed. A
hlgher-prlorltv device can Interrupt a lower-priority service routine without any
difficulty. Note that a RETI Instruction at the end of the high Priority routine will not be
recognized by the lower-priority device.

SIO INTERRUPTS
The 510 can also serve as a source for interrupts. You should note
the follOWing features of the 510 Interrupt-based systems:
1)
2)

3)
4)

SIO
INTERRUPTS

The transmitter Interrupt IS enabled by setting bit 1 of Write

Register 1 on each channel.
The Interrupt vector is affected by bits 2. 3. and 4 of Write Register 1 according to
Tables 12-2 and 12-3.
The Interrupt vector IS In Write Register 2 on Channel B only. It can be read from
Read Register 2 on Channel B only.
Bit 01 of Read Register 0 on Channel A IS 1 if any Interrupt condition is presenH1j1
the entire 510.
,.

Within an 510. Channel A Interrupts have priority over Channel B interrupts. receiver Inl
terrupts have priOrity over transmitter interrupts. and transmitter Interrupts have
priOrity over external or status interrupts.
SIOs can be used in a polling Interrupt system. The CPU must
check each 510 for activity by examining bit 1 of Read Register 9
on Channel A. I.e.,
SUB
OUT
IN
BIT

JR

A
(SIOCRAl.A
A.(SIOCRA)
1.A
NZ.SERVE

The Important features of a

1)

2)

:ACCESS READ REGISTER 0

POLLING
INTERRUPT
SYSTEMS
WITH SIOs

;GET 510 STATUS

:ANY INTERRUPTS PENDING?
;YES. INTERRUPT ACTIVE

zao polling system are:

The first Interrupt examined has the highest priority. since the remaining interrupts
will not be examined if the first one is active. The second interrupt has the next
highest priOrity. and so on.
The service routine must clear the SIO Interrupt by reading or writing the appropriate data register even if a data transfer is otherwise unnecessary.

12-10

Table 12-2. Further Vectoring of SIO Interrupts

(Bit 2 of SIO Write Register 1 on Channel B is 1)
Status Affects Vector (02) (Channel B Onlvl
if this bit

IS

1. the vector retumed from an Interrupt acknowtedge cycle will be venable according to the following:

Ch B

Ch A

V3

V2

Vl

0
0
0
0

0
0
1

Ch 8 Transmit Buffer Empty

Ch B ExtemaJfStatus Change

Ch B Receive Character Available

Ch B Special Receive Condition-

0
0
1
1

Ch A Transmit Buffer Empty

Ch A External/Status Change

0
1

Ch A Special Receive Condition-

1
1

Ch A Receive Character Available

Speclal Receive Conditions .....

Panty Error or
Rx Overrun Error or
CRC/Framing Error or

End of Frame (SOLCl

If this bit is O. the fixed vector programmed in the Interrupt Vector register is returned.

Table 12-3. SIO Interrupt Modes

(Bits 3 and 4 of Write Register 1)
Rec Int Mode 0 (03). Rec Int Mode 1 (04)

Receive Interrupt Mode 0 and Receive Interrupt Mode 1 together specify the venous character available conditions:

Mode

0
1

04

03

Rec Int
Mode 1

Rec Int

0
0

Mode 0
0
1

Receiver interrupts disabled

Receive Interrupt on first character
onlverror

Interrupt on all Receive CharactersParity error affects Vector

Interrupt on all Receive CharactersParity error does not affect Vector

12-11

INTERRUPT EXAMPLES
A Startup Interrupt
Purpose: The computer waits for a Pia interrupt to occur before starting actual operations.
Many systems remain inactive until the operator actually starts them or a DATA READY
signal IS received. On RESET. such systems must initialize the Stack Pointer. enable the
startup interrupt. and execute a HALT Instruction. Remember that RESET disables the
processor interrupt and power-on disables all Pia Interrupts. In the flowchart. the decision as to whether startup is active is made In hardware O.e.. by the CPU examining the
interrupt Input Internally) rather than in software.
Flowchart:

Initialize Stack Pointer

Enable stertup
mterrupt on PIO

Enable CPU interrupt

Source Program:
Main Program:
RESET

EQU
ORG
LD
LD
OUT
LD
OUT
EI
HALT

o
RESET
SP.l00H
A.Ol 001111 B
(PIOCRA).A
A.l 0000111 B
(PIOCRA).A

:PUT STACK AT END OF MEMORY

:PUT Pia IN INPUT MODE
:ENABLE Pia INTERRUPT
:ENABLE INTERRUPTS
:AND WAIT

Interrupt Service Routine:

ORG
LD

JP

INTRP
SP.l00H
START

:REINITIALIZE STACK POINTER

;START MAIN PROGRAM

12-12

Object Program:

Memory Address
(Hex)

Memory Contents
(Hexl

Main Program:
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC

Instruction
(Mnemonic)

31
00
01
3E
4F
03
PIOCRA
3E
87
03
PIOCRA
FB
76

LD

SP.100H

LD

A.01001111 B

OUT

(PIOCRA),A

LD

A.10000111 B

OUT

(PIOCRA),A

31
00
01
C3

LD

SP.100H

JP

START

EI
HALT

Interrupt Service Routine:

INTRP
INTRP+1
INTRP+2
INTRP+3
INTRP+4
INTRP+5

START

The main program must initialize the Stack Pointer. since the interrupt response always
stores the old Program Counter in the Stack. Here the service routine simply reinltializes
the Stack Pointer before the actual startup routine is executed. An alternative would be
to increment the Stack POinter twice before jumping to the startup routine. Remember
that the Z80 comes up In Interrupt Mode O. Any other mode would require the execution of an 1M Instruction.
INTERRUPTS ON
The exact location of the Interrupt service routine varies
PARTICULAR
with the microcomputer. If your microcomputer has no
MICROCOMPUTERS
monitor. you can start the interrupt service routine
wherever the external hardware or vector table directs the
CPU. Of course. yOU should place the routine so that It does not Interfere with fixed addresses or with other programs.

If your microcomputer has a monitor. the monitor will often occupy the RESET and Interrupt service addresses. It will then supply
service routines or the addresses of those routines. A typical monitor routine initialization would be:
.
MONIN:

PUSH
LD
JP

HL
HL.USRINT
(HU

INTERRUPT
HANDLING
BY
MONITORS

;SAVE OLD REGISTER CONTENTS

;GET USER ADDRESS FOR SERVICE
;JUMP TO USER SERVICE ADDRESS

You must then place the address of your service routine into memory locations USRINT
and USRINT+1. using the normal Z80 address format with the least significant bits at
the lower address. Remember that MONIN is an address in the monitor program.

12-13

You can Include the loading of memory locations USRINT and USRINT+1 in your main
program: I.e..
LD

HUNTRP

LD

(USRINTl.HL

:GET STARTING ADDRESS OF SERVICE

ROUTINE
:STORE IT AS USER ADDRESS

These instructions come before the enabling of the Interrupts.

In this example. the return address that the zao stores in the Stack is not useful.
However. the main program still must initialize the Stack POinter so that there is a
definite place to put that address. You may not need the LD SP instruction if the monitor In your microcomputer manages the Stack POinter.
The main program enables onlY the interrupt from the startup Pia. The Pia could, of
course, be In any mode. The interru pt is enabled by setting bit 7 of an Interrupt control
word and writing that word to the Pia control port. The Pia interrupt is enabled before
the overall Interrupt system IS enabled With the EI instruction.
Remember that RESET and accepting an interrupt automatically disable the Interrupt
system. This allows the real startup routine to configure all the PIOs and other Interrupt
sources without being Interrupted.
No action IS needed In the interrupt service routine, since the Interrupt IS automatically
cleared as part of the Interrupt Acknowledge cycle Involving a particular Pia.
The implementations of the instructions EI (Enable Interrupts) and 01 (Disable Interrupts) differ on the zao. 01 takes effect Immediately after its execution, while EI takes
effect after the execution of the following Instruction. The reasoning behind thiS fact is
discussed In Chapter 3 under the deSCription of the EI Instruction.

A Keyboard Interrupt
Purpose: The computer waits for a keyboard interrupt and places
the data from the keyboard into memory location 0040.
Sample Problem:

Keyboafd data
Result:

(0040)

06
06

Flowchart:

Initialize Stack POInter

Enable kevDoaro
Interrupt on rJlO
Enable ~PU interrupt

1401

data

12-14

Source Program:

Main Program:
RESET

HERE:

EQU
ORG
LD
LD
OUT
LD
OUT
EI
JR

o
RESET
SP,lOOH
A.01001111B
(PIOCRA).A
A,10000111B
(PIOCRA).A

:PUT STACK AT END OF MEMORY

:PUT PIO IN INPUT MODE
;ENABLE PIO INTERRUPTS
;ENABLE CPU INTERRUPTS
;DUMMY MAIN PROGRAM

HERE

Interrupt Service Routine:

ORG
EX
IN
LD
EX
EI
RETI

INTRP
AF.AF'
A.(PIODRA)
(40H).A
AF.AF

;SAVE ACCUMULATOR. FLAGS

;GET KEYBOARD DATA
;SAVE KEYBOARD DATA
;RESTORE ACCUMULATOR. FLAGS
:RE-ENABLE INTERRUPTS

Object Program:

Memorv Address
(Hex)

Memorv Contents
(Hex)

Instruction
(Mnemonic)

Main Program:
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
0008
OOOC
OOOD

31
00
01
3E
4F
D3
PIOCRA
3E
87
D3
PIOCRA
FB
18
FE

HERE:

LD

SP,100H

LD

A.01001111 B

OUT

(PIOCRA).A

LD

A,10000111B

OUT

(PIOCRA).A

EI
JR

HERE

EX
IN

AF.AF'
A.(PIODRA)

LD

(40Hl.A

EX
EI
RETI

AF,AF'

Interruot Service Routine:

INTRP
INTRP+1
INTRP+2
INTRP+3
INTRP+4
INTRP+5
INTRP+6
INTRP+7
INTRP+8
INTRP+9

08
DB
PIODRA
32
40
00
08
FB
ED
4D

12-15

The JR HERE is an endless loop (jump-ta-self) instruction that is used to represent the
main program. After interrupts are enabled in a working system, the main program goes
about its business until an Interrupt occurs and then resumes execution after the interrupt service routine is completed.
The RET instruction at the end of the service routine transfers
control back to the JR instructiOn. If you want to avoid this, you
can simply Increment the Program Counter In the Stack, e.g..
EX
INC
INC
EX

(SP).HL
HL
HL
(SPl.HL

CHANGING THE
RETURN
ADDRESS

:GET RETURN ADDRESS

:INCREMENT RETURN ADDRESS TWICE
:RESTORE ADJUSTED ADDRESS TO STACK

The RET instruction will now transfer control to the Instruction follOWing the JR. Note
the use of EX (SP).HL: this instruction exchanges the contents of Register Pair HL with
the contents of the memory locatiOns at the top of the Stack. By using It we can adjust
the return address without affecting the contents of Register Pair HL.
Since the zao does not automatically save its registers, you can use them to pass
parameters and results between the main program and the interrupt service routine. So,
yOU could leave the data in the Accumulator instead of in memory location 0040. This
is, however, a dangerous practice that should be aVOided in all but the most triVial
systems. In most applications, the processor is using its registers during normal program execution: having the Interrupt service routines randomly change the contents of
those registers would surely cause havoc. In general. no Interrupt service routine should
ever alter any register unless that register's contents have been saved prior to its alteration and will be restored at the completion of the routine.
Note that you must explicitly re-enable the Interrupts anhe end of the service routine,
since the processor disables the interrupt system when It accepts an interrupt. Servicing a PIO interrupt deactivates the interrupt Signal so that the same Interrupt is not serviced again.
SAVING
If interrupt service routines are never themselves interrupted (i.e..
there is only one level of interrupts). the instructions EX AFAF
VALUES IN
and EXX are a convenient way to save and restore the old contents
PRIMED
of the user registers. EXX exchanges the contents of BC, DE, and
REGISTERS
HL with the contents of their primed equlyalents. The two instructions together take only two bytes of memory and eight clock cycles. However, this
method cannot be used if there are other interrupt levels (since there IS only a single set
of primed registers) or if the primed registers are needed In either the main program or
the Interrupt service routine.

A more general approach to saving and restoring registers is to use the Stack. PUSH
sayes the contents of a register pair and POP restores the contents. Howeyer, PUSH
takes 11 clock cycles and POP 10, so this approach is slower. It also uses extra memorv
locations In the Stack. The advantage of this method IS that it can be expanded indefinitely (as long as there is room in the Stack) since nested service routines will not
destroy the data saved by the earlier routines.
An alternative approach would be for the Interrupt routine to
maintain control until it received an entire line of text (e.g.. a string
of characters ending with a camage return). The main program
would be:

12-16

FILLING A
BUFFER VIA
INTERRUPTS

Main Program:
RESET

HERE:

EQU
ORG
LD
LD
OUT
LD
OUT
LD
LD
EI
JR

0
RESET
SP,100H
A.01001111 B
(PIOCRA),A
A.10000111 B
(PIOCRA),A
HL.70H
(40Hl.HL
HERE

;PUT STACK AT END OF MEMORY

:PUT PIO IN INPUT MODE
;ENABLE PIO INTERRUPTS
;INITIALIZE BUFFER POINTER
;SAVE BUFFER POINTER
;ENABLE CPU INTERRUPT
:DUMMY MAIN PROGRAM

Interrupt Service Routine:

ENDL:

ORG
EX
EXX
LD
IN
LD
CP
JR
INC
LD
EXX
EX
EI
RETI
JP

INTRP
AF,AF
HL.(40H)
A.(PIODRA)
(HL).A
CR
Z,ENDL
HL
(40Hl.HL

;SAVE A. FLAGS
:SAVE OTHER REGISTERS
:GET BUFFER POINTER
;GETKEYBOARD DATA
:SAVE DATA IN BUFFER
:IS DATA A CARRIAGE RETURN?
:YES. END OF LINE
:NO. INCREMENT BUFFER POINTER

AF.AF

;RESTORE OTHER REGISTERS

:RESTORE A. FLAGS
;RE-ENABLE INTERRUPTS

LPROC

;PROCESS LINE WITHOUT INTERRUPTS

When the processor receives a carriage return. It leaves the Interrupt system disabled
while It handles the line,
An alternative approach would be to fill another buffer while handling the first one: this approach is called double buffering.
The line processing routine IS begun at address LPROC with interrupts disabled. the old register contents in the primed registers. and the onglnal return
address at the top of the Stack.
In a real application. the CPU could perform other tasks between interrupts. It could. for
Instance. edit. move. or transmit a line from one buffer while the Interrupt was filling
another buffer.

A Printer Interrupt
Purpose: The computer waits for a printer interrupt and sends the data from memory
location 0040 to the printer.
Sample Problem:
(0040)

51H

Result:

Pnnter receives a 51 H (ASCII Qj when It is ready,

12-17

Flowchart:

Initialize Stack Pointer

Enable pnnter
interrupt on PIO
Data =(40)

Enabla CPU interrupt

Send data to prjnter

Source Program:
Main Program:
RESET

HERE:

EQU
GRG
LD
LD
OUT
LD
OUT
EI
JR

a
RESET
SP,100H
A.00001111 B
(PIOCRA),A
A.1 0000111 B
(PIOCRA),A
HERE

;PUT STACK AT END OF MEMORY

:PUT PIO IN OUTPUT MODE
;ENABLE PIO INTERRUPTS
;ENABLE CPU INTERRUPTS
;DUMMY MAIN PROGRAM

Interrupt Service Routine:

ORG

EX
LD
OUT

EX
EI
RETI

INTRP
AF.AF
A. (40H)
(PIODRA),A
AF.AF

;SAVE ACCUMULATOR, FLAGS

:GET DATA
;SEND DATA TO PRINTER
;RESTORE ACCUMULATOR. FLAGS
:RE-ENABLE INTERRUPTS

12-18

Object Program:
Memorv Address
(Hex)

Memorv Contents
(Hex)

Instruction
(Mnemonic)

Main Program:
0000
0001
0002
0003
0004
0005
0006
0007
OOOS
0009
OOOA
OOOB
OOOC
0000

31
00
01
3E
OF
03
PIOCRA
3E
S7
03
PIOCRA
FS
lS
FE

HERE:

LD

SP.l00H

LD

A.OOOOllll B

OUT

(PIOCRA).A

LD

A.l0000111 B

OUT

(PIOCRA).A

EI
JR

HERE

EX
lD

AF.AF
A.(40H)

OUT

(PIODRA).A

EX
EI
RETI

AF.AF

Interrupt Service Routine:

OS
3A
40
00
03
PIODRA
OS
FB
ED
40

INTRP
INTRP+l
INTRP+2
INTRP+3
INTRP+4
INTRP+5
INTRP+6
INTRP+7
INTRP+S
INTRP+9

Here. as with the keyboard. you could have the printer continue to
interrupt until it transferred an entire line of text. The main program and the service routine would be:

EMPTYING A
BUFFER WITH
INTERRUPTS

Main Program:
RESET

HERE:

EQU
ORG
LD
LD
OUT
lD
OUT
LD
LD
EI
JR

o
RESET
SP.l00H
A.00001111 B
(PIOCRA).A
A.10000111 B
(PIOCRA),A
HL,70H
(40H).HL
HERE

:PUT STACK AT END OF MEMORY

:PUT PIO IN OUTPUT MODE
;ENABLE PIO INTERRUPTS
:INITIALIZE BUFFER POINTER
SAVE SUFFER POINTER
ENABLE CPU INTERRUPT
DUMMY MAIN PROGRAM

12-19

Interrupt Service Routine:

ENOL:

ORG
EX
EXX
LD
LD
OUT
CP
JR
INC
LD
EXX
EX
EI
RETI
JP

INTRP
AF.AF
HL.(40H)
A.(HU
(PIODRAl.A
CR
ENOL
HL
(40Hl.HL

;SAVE A. FLAGS
;SAVE OTHER REGlSTERS
:GET BUFFER POINTER
:GET A BYTE OF DATA FROM BUFFER
:SEND DATA TO PRINTER
:IS DAtA A CARRIAGE RETURN?
;YES. END OF LINE
:NO. INCREMENT BUFFER POINTER

AF.AF'

:RESTORE OTHER REGISTERS

:RESTORE A. FLAGS
:RE-ENABLE INTERRUPTS

LCOMP

:HANDLE COMPLETED LINE

Again. double buffering could be used to allow I/O and processing to occur at the same
time without ever halting the CPU

A Real-Time Clock Interrupt

Purpose: The computer waits for an interrupt from a real-time
clock.

A real-time clock simply provides a regular series of pulses. The interval between the pulses can be used as a time reference. Real-time clock interrupts
can be counted to give any multiple of the baSIC time Interval. A real-time clock can be
produced by dividing down the CPU clock. by uSing a separate timer or a programmable timer like the CTC for ZBO-based microcomputers. or by using external sources such
as the AC line frequency.
FREQUENCY
Note the tradeoffs involved in determining the frequency of the
real-time clock. A high frequency (say 10 kHz) allows the creaOF REAL-TIME
CLOCK
tion of a wide range of time intervals of high accuracy. On the
other hand. the overhead Involved In counting real-time clock
Interrupts may be considerable. and the counts will quickly exceed the capacity of a
single B-bit register or memory location. The choice of frequency depends on the precIsion and timing requirements of your application. The clock may. of course. consist
partly of hardware: a counter may count high frequency pulses and Interrupt the processor only occasionally. A program will have to read the counter to measure time to
high accuracy.

One problem IS synchronizing operations with the real-time

clock. Clearly. there will be some effect on the precision of
the timing interval if the CPU starts the measurement randomly dUring a clock period. rather than exactly at the
beginning. Some ways to synchronize operations are:
1)
2)

SYNCHRONIZATION
WITH REAL-TIME
CLOCK

Start the CPU and clock together. RESET or a startup interrupt can start the clock as
well as the CPU.
Allow the CPU to start and stop the clock under program control.

3)

Use a high-frequency clock so that an error of less than one clock period will be
small.

4)

line up the clock (by waiting for an edge or interrupt) before starting the measurement.

12-20

A real-time clock interrupt should have very high priority, since

the precIsion of the timing Intervals will be affected by any delay
In servicing the interrupt. The usual practice IS to make the realtime clock the highest Priority Interrupt except for power failure.
The clock interrupt service routine is generally kept extremely short
interfere with other CPU activities.

a)

PRIORITY
OF REAL-TIME
CLOCK
so that it does not

Wait for Real-Time Clock

Source Program:
Main Program:
RESET

HERE:

EQU
ORG
LD
LD
OUT
LD
OUT
EI
JR

o
RESET
SP.100H
A.01001111B
(PIOCRAl.A
A,10000111B
(PIOCRAl.A

:PUT STACK AT END OF MEMORY

:PUT PIO IN INPUT MODE
:ENABLE PIO INTERRUPTS
:ENABLE CPU INTERRUPTS
:DUMMY MAIN PROGRAM

HERE

Interrupt Service Routine:

ORG
HALT

INTRP
;END CLOCK INTERRUPT

Object Program:
Memory Address
(Hex)

Instruction
(Mnemonic)

Memory Contents
(Hex)

Main Program:
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009

OOOA
OOOB
OOOC
OOOD

31
00
01
3E
4F
D3
PIOCRA
3E
87
D3
PIOCRA
FB
18
FE

HERE:

LD

SP,100H

LD

A,01001111B

OUT

(PIOCRAl.A

LD

A.10000111B

OUT

(PIOCRAl.A

EI
JR

HERE

Interrupt Service Routine:

INTRP

HALT

76

The service routine does not have to do anything, since servicing the PIO Interrupt automatically clears It and there is no data to send or receive.
The real-time clock interrupt always occurs on'a rising edge if a PIO STROBE signal is
used for the clock input.

12-21

bl

Wait for 10 Real-Time Clock Interrupts

Source Program:

Main Program:
RESET

EQU
ORG
LD
LD
OUT
LD
OUT
LD
LD
LD
EI
WTTEN: CP

JR

o
RESET
SP,100H
A0100llllB
(PIOCRAl.A
A.l0000lllB
(PIOCRAl.A
HL.40H
(HL).O
Al0
(HL)
NZ.wTTEN

HALT

:PUT STACK AT END OF MEMORY

:PUT PIO IN INPUT MODE
:ENABLE PIO INTERRUPTS
:CLOCK COUNTER

= ZERO

:NUMBER OF COUNTS = 10
:ENABLE CPU INTERRUPTS
:HAVE TEN COUNTS ELAPSED?
:NO. WAIT
:YES. DONE

Interrupt Service Routine:

ORG

EXX
EX
LD
INC

EX
EXX
EI
RETI

INTRP
AF.AF'
HL.40H
(HL)
AF.AF'

:SAVE USER REGISTERS

:SAVE A FLAGS
:INCREMENT CLOCK COUNTER
:RESTORE A. FLAGS
:RESTORE USER REGISTERS
:RE-ENABLE INTERRUPTS

12-22

Object Program:
Memory Address
(Hex)

M!lmory Contents
(Hex)

Instruction
(Mnemonic)

Main Program:
0000
0001
0002
0003
0004
0005
0006
0007
OOOS
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016

31
00
01
3E
4F
D3
PIOCRA
3E
S7
D3
PIOCRA
21
40
00
36
00
3E
OA
FB
BE
20

WTTEN:

LD

SP.l00H

LD

A.Ol001111 B

OUT

(PIOCRAl.A

LD

A,10000111B

OUT

(PIOCRAl.A

LD

HLAOH

LD

(HL).O

LD

A.l0

EI
CP
JR

(HL)
NZ.WTTEN

FD
76

HALT

U9

EXX
EX
LD

Interrupt Service Routine:

INTRP
INTRP+l
INTRP+2
INTRP+3
INTRP+4
INTRP+5
INTRP+6
INTRP+7
INTRP+S
INTRP+9
INTRP+l0

OS
21
40
00
34
OS
D9
FB
ED
4D

INC
EX
EXX
EI
RETI

AF.AF'
HLAOH

(HL)
AF.AF'

An alternative approach uses the Stack to save and restore register values. To save H.
and the flags requires:
PUSH
PUSH

HL
AF

L.

:SAVE REGISTERS HAND L

:SAVE ACCUMULATOR AND FLAGS

To restore them requires the sequence:

POP
POP

AF
HL

:RESTORE ACCUMULATOR AND FLAGS

:RESTORE REGISTERS HAND L

Note that. if the Stack IS used. registers must be restored In the opposite order from that
which they were saved. Clearly the order In which EXX and EX AF.AF' are executed
does not matter.
In

12-23

This Interrupt service routine merely updates the counter In memorv location 0040. It
transparent to the main program.

A more realistic real-time clock interrupt routine could

maintain real time in several memory locations. For example. the following routine uses addresses 0040 through 0043
as follows:
0040
0041
0042
0043

hundredths of seconds
seconds
minutes
hours

We assume that the routine

Flowchart:

IS

triggered by a 100 Hz clock.

Clear clock interrupt

Hundredths =
Hundredths + 1

Hundredths
Seconds

Seconds + 1

Seconds = 0
Minutes
Minutes + 1

Minutes
Hours

Hours

End

12-24

+ 1

IS

MAINTAINiNG
REAL TIME

Source Program:

DONE:

ORG
PUSH
PUSH
LD
INC
LD
CP
JR
LD
INC
INC
LD
CP
JR
LD
INC
INC
CP
JR
LD
INC
INC
LD
JR
LD
POP
POP
EI
RETI

INTRP
AF
HL
HLAOH
(HL)
A.100
(HL)
NZ.DONE
(HL).O
HL
(HL)
A.50
(HL)
NZ.DONE
(HL).O
HL
(HL)
(HL)
NZ,DONE
(HL).O
HL
(HL)
A.24
NZ,DONE
(HL).O
HL
AF

:SAVE REGISTERS
:UPDATE HUNDREDTHS OF SECONDS

;IS THERE A CARRY TO SECONDS?

:NO. DONE
;YES. HUNDREDTHS = 0
:UPDATE SECONDS

:IS THERE A CARRY TO MINUTES?

:NO. DONE
;YES. SECONDS = 0
:UPDATE MINUTES
:IS THERE A CARRY TO HOURS?
:NO. DONE
:YES, MINUTES = 0
:UPDATE HOURS
:DAY COMPLETED?
:NO. DONE
:YES, HOURS = 0
:RESTORE REGISTERS
:RE-ENABLE INTERRUPTS

Now a walt of 300 ms could be produced in the main program with the routine:

WT30:

LD
LD
ADD
CP
JR
SUB
CP
JR

HLAOH
A.(HL)
A.30
100

:DESIRED TIME IS 30 COUNTS LATER

;MOD 100

100
(HL)
NZ,WT30

:DESIRED TIME REACHED?

;NO, WAIT

c,Wno

:GET PRESENT TIME (HUNDREDTHS OF SECS)

Be careful In thiS program of the difference between INC HL and INC (HL). INC HL adds
1 to the 15-blt contents of Register Pair HL. while INC (HU adds 1 to the B-bit contents
of the memory location addressed by HL.
Of course, the program could perform other tasks and check the elapsed time only occasionally. How would you produce a delay of seven seconds? Of three minutes?
Sometimes you may want to keep time either as BCD digits or as ASCII characters. How
would you revise the last program to handle these alternatives?
You can disable the clock Interrupt (or any other interrupt) when it
IS no longer needed In any of the following ways.

DISABLING
INTERRUPTS

1)

By executing a 01 instruction in the main program. This disables the entire interrupt system.

2)

By clearing bit 7 of the interrupt control word dUring the service routine or dUring
the main program. This disables only the interrupt from one port of one PIO.

12-25

3)

By not re-enabling the interrupt during' the service routine.

Remember that the CPU automatically disables interrupts upon accepting one. Thus.
the interrupt system is disabled unless the service routine explicitly re-enables it. Note.
however. that you must be very careful about not re-enabling the interrupts. since the
main program would be completely unaware that interrupts were no longer allowed. In
general. all Interrupt service routines should 're-enable the interrupts before returning; any other policy means that the service routines are not transparent to the main
program.

A Teletypewriter Interrupt
Purpose: The computer walts for data to be received from a teletypewriter and stores
the data in memory location 0040.
al

Using an SIO

(7-bit characters with odd parity and 2 stop bits).

SIO
INTERRUPT
ROUTINE

Source Program:
Main Program:
RESET

HERE:

EQU
LD
OUT
LD
OUT
LD
OUT
LD
OUT
LD
OUT
LD
OUT
EI
JR

0
AA
(SIOCRA).A
A01000001B
{SIOCRA).A
A.3
{SIOCRA).A
A.01000001B
{SIOCRA).A
A1
(SIOCRA).A
A.00011000B
(SIOCRA).A
HERE

:ACCESS WRITE REGISTER 4

;X16 CLOCK MODE. PARITY

;7 BIT CHARACTERS. ENABLE RECEIVER

;ACCESS WRITE REGISTER 1
;ENABLE RECEIVER INTERRUPT ON ALL CHARS
;ENABLE CPU INTERRUPTS
;DUMMY MAIN PROGRAM

Interrupt Service Routine:

ORG
PUSH
IN
LD
POP
EI
RETI

INTRP
AF
A.(SIODRA)
(40H).A
AF

;SAVE ACCUMULATOR. FLAGS

;READ CHARACTER FROM SIO
;SAVE CHARACTER IN MEMORY
;RESTORE ACCUMULATOR. FLAGS
;RE-ENABLE INTERRUPTS

12-26

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

Instruction
(Mnemonic)

Main Program:
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
001A

3E
04
D3
SIOCRA
3E
41
D3
SIOCRA
3E
03
D3
SIOCRA
3E
41
D3
SIOCRA
3E
01
D3
SIOCRA
3E
18
D3
SIOCRA
FB
18
FE

HERE:

LD

AA

OUT

(SIOCRA).A

LD

A01000001B

OUT

(SIOCRA).A

LD

A.3

OUT

(SIOCRAl.A

LD

AOl 000001 B

OUT

(SIOCRA),A

LD

A.l

OUT

(SIOCRAl.A

LD

A.00011000B

OUT

(SIOCRA).A

EI
JR

HERE

PUSH
IN

AF
A.(SIODRA)

LD

(40Hl.A

POP
EI
RETI

AF

Interrupt Service Routine:

INTRP
INTRP+l
INTRP+2
INTRP+3
INTRP+4
INTRP+5
INTRP+6
INTRP+7
INTRP+8
INTRP+9

F5
DB
SIODRA
32
40
00
Fl
FB
ED
4D

This service routine assumes that only the receive interrupt from one channel of the SIO
has been enabled. Otherwise. either further vectoring will be reqUired by changing control bits 02. 03. and 04 of Write Register 0 (see the discussion of SIO interrupts earlier
in this chapter) or the routine will have to examine the status bits in Read Register O.
The key status bits are:
Bit 0 - Receive Character Available -1 when at least one character is available in the
receive buffers.
Bit 1 - Interrupt pending (Channel A only) -1 if any interrupt IS pending in the entire
SIO.

12-27

Bit 2 - Transmit Buffer Empty -1 if the Transmit buffer IS empty.

Obviously. It would be far shorter and simpler to configure the SIO by using a table (in
ROM) and the repeated Block I/O Instruction, i.e..
LD
LD
LD
OTIR

B.6
C,SIOCRA
HL.SIOTBL

:NUMBER OF BYTES IN CONFIGURATION

;SIO CONTROL PORT
:START OF SIO CONFIGURATION TABLE
;CONFIGURE SIO

This method requires 9 bytes of memory for the program and 6 bytes for the table, as
compared to the 23 bytes used In the example to configure the SIO.
The program establishes the SIO registers as follows:
WRITE REGISTER 4
Bit 7

= 0,

bit 6

= 1 for X16

clock mode

Bit 1 = 0 to select odd panty

Bit 0 = 1 to enable panty generation
WRITE REGISTER 3
Bit 7 = 0, bit 6 = 1 to select 7-bit characters
Bit 0 = 1 to enable the receiver
WRITE REGISTER 1
Bit 4 = 1, bit 3 = 1 to produce an Interrupt on all received characters With parity errors
not affecting the vector.
The CPU clears the Received Character Available bit by reading a character from the
SIO Data register. The Interrupt Pending bit IS cleared automatically when the Interrupt
is serviced.

bl

Using a PIO

(Received data tied to data bit 7 of PIO Port A),

Source Program:
Main Program:

HERE:

LD
OUT
LD
OUT
LD
OUT
LD
OUT
EI
JR

A,11001111B
(PIOCRA),A
A,10000000B
(PIOCRAl.A
A,10010111B
(PIOCRA).A
A,01111111B
(PIOCRAl.A
HERE

;MAKE PORT A CONTROL

:MAKE BIT 7 INPUT, OTHERS OUTPUTS
:ENABLE INTERRUPT ON START BIT (0)
:MASK OUT ALL OTHER BITS
:ENA8LE CPU INTERRUPTS
:DUMMY MAIN PROGRAM

Interrupt Service Routine:

ORG
PUSH
LD
OUT
CALL
LD
OUT
POP
EI
RETI

INTRP
AF
A,000001118
(PIOCRAl.A
TTYRCV
A,100001118
(PIOCRAl.A
AF

:SAVE ACCUMULATOR, FLAGS

:DISABLE START 81T INTERRUPT
:FETCH DATA FROM TTY
:ENABLE START BIT INTERRUPT
:RESTORE ACCUMULATOR, FLAGS
:RE-ENABLE INTERRUPTS

12-28

Object Program:
Memory Address
(Hex)

Memory Contents
(Hex)

Instruction
(Mnemonic)

Main Program:
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
OOOA
OOOB
OOOC
0000
OOOE
OOOF
0010
0011
0012

LO

A,11 001111 B

OUT

(PIOCRAl.A

LO

A, 1OOOOOOOB

OUT

(PIOCRAl.A

LO

A,10010111B

OUT

(PIOCRA).A

LO

A.Olll1111 B

OUT

(PIOCRAl.A

EI
JR

HERE

F5
3E
07
03
PIOCRA
CO
TIVRCV

PUSH
LO

AF
A.OOOOOlll B

OUT

(PIOCRA).A

CALL

TIVRCV

3E
87
03
PIOCRA
Fl
FB
EO
40

LO

A.l0000lllB

OUT

{PIOCRA).A

POP
EI
RETI

AF

3E
CF
03
PIOCRA
3E
80
03
PIOCRA
3E
97
03
PIOCRA
3E
7F
03
PIOCRA
FB
18
FE

HERE:

l'lterrupt Service Routine:

INTRP
INTRP+l
INTRP+2
INTRP+3
INTRP+4
INTRP+5
INTRP+6
INTRP+7
INTRP+8
INTRP+9
INTRP+l0
INTRP+ll
INTRP+12
INTRP+13
INTRP+14
INTRP+15

These programs assume that the monitor initializes the Stack Pointer. Otherwise. it will
have to be loaded In the main program.
Subroutine TTVRCV is the TTV receive routine shown In the prevIous chapter.
The edge used to cause the Interrupt IS very Important here. An Interrupt must occur
when the data line changes from the normal MARK or '1' state to the SPACE or '0' state.
since this transition identifies the start of the transmission.
The service routine must disable the PIO Interrupt. since otherwise each '1' -to-'O' transition In the character will cause an interrupt. Of course. you must re-enable the PIO interrupt after the entire character has been read.

12-29

Note the use of the PIO In the control mode:

1)
2)

3)

4)

The PIO is placed in the control mode by establishing Mode 3.

The next control word defines which data lines are to be inputs ('1') and which are
to be outputs ('0').
The interrupt control word has. besides the usual enable In bit 7,
bit 6 = 0 to perform a logical OR of the monitored data lines for an interrupt (not
used in this case. since onlY one line is monitored)
bit 5 = 0 to define the actlye polarity of the data lines as low (for the start bit in thiS
case)
bit 4 = 1 to indicate that a mask word follows.
The next control word contains the Interrupt masks. Only those port lines with a
mask bit of zero will be monitored for generating an Interrupt.

The net result is for an interrupt to be generated if bit 7 is zero or changes from one to
zero. Note that further interrupts occur only when a change occurs In the status of the
logical equation. Here again. the PIO could be configured by using a table and the repeated block output Instruction.

MORE GENERAL SERVICE ROUTINES

More general service routines that are part of a complete interrupt-driven system must handle the follOWing tasks:
1)

Saving all registers that are used In the Interrupt serVice

routine In the Stack so that the Interrupted program can be
correctly resu med.

TASKS FOR
GENERAL SERVICE
ROUTINES

Remember that the Z80 Push instruction transfers a register pair (or an index register)
to the Stack. PUSH AF (F is the Flag register) transfers the Accumulator and flags to the
Stack.

A routine to save all the registers In the Stack would be:

PUSH
PUSH
PUSH
PUSH
PUSH
PUSH
EX
EXX
PUSH
PUSH
PUSH
PUSH

AF
BC
DE
HL
IX
IY
AF.AF'

:SAVE
:SAVE
;SAVE
:SAVE
:SAVE
;SAVE

ACCUMULATOR. FLAGS
REGISTERS B.C
REGISTERS D.E
REGISTERS H.L
INDEX REGISTER IX
INDEX REGISTER IY

AF
BC
DE
HL

;SAVE
;SAVE
;SAVE
;SAVE

PRIMED
PRIMED
PRIMED
PRIMED

ACCUMULATOR. FLAGS
REGISTERS B.C
REGISTERS D.E
REGISTERS H.L

Of course. only those registers that are used by the Interrupt service routine must be
saved.
2)

3)

RestOring all registers from the Stack after completing the Interrupt service routine.
Remember that registers must be restored In the opposite order from that In which
they were saved.
Enabling and disabling interrupts appropriately. Remember that the CPU automatically disables ItS Interrupts upon accepting one.

The service routines should be transparent as far as the Interrupted program IS concerned (j.e.. they should have no incidental effects).

12-30

Any standard subroutines that are used by an Interrupt service routine must be
reentrant. If some subroutines cannot be made reentrant. the interrupt service routine
must have separate versions to use. 5

PROBLEMS
1) A Test Interrupt
Purpose: The computer walts for a Pia interrupt to occur. then executes the endless
loop instruction:

HERE:

JR

HERE

until the next Interrupt occurs.

2)

A Keyboard Interrupt

Purpose: The computer waits for a 4-digit entry from a keyboard and places the digits
Into memory locations 0040 through 0043 (first one received in 0040). Each
digit entry causes an Interrupt. The fourth entry should also result in the disabling of the keyboard interrupt.
Sample Problem:

Keyboard data = 04. 06, 01. 07

Result:

3)

(0040)
(0041)
(0042)
(0043)

04
06
01
07

A Printer Interrupt

Purpose: The computer sends four characters from memory locations 0040 to 0043
(starting with 0040) to the printer. Each character is requested by an interrupt. The fourth transfer also disables the printer interrupt.

4)

A Real-Time Clock Interrupt

Purpose: The computer clears memory location 0040 initially and then complements
memory location 0040 each time the real-time clock Interrupt occurs.

How would you change the program so that it complements memory location 0040
after every ten Interrupts? How would you change the program so that it leaves memory location 0040 at zero for ten clock penods. FF16 for five clock periods, and so on continuously? You may want to use a display rather than memory location 0040 so that it
will be easier to see.

5)

A Teletypewriter Interrupt

Purpose: The computer receives TTY data from an interrupting SIO and stores the
characters In a buffer starting in memory location 0040. The process continues until the computer receives a carriage return (OD16).

Assume that the characters are 7-blt ASCII With odd panty. How would you change
your program to use a Pia? Assume that subroutine TTYRCV is available. as In the example. Include the carriage return as the final character in the buffer.

12-31

REfERENCES
1.

You may want to review the discussion of Interrupts In Volume 1 of An Introduction

to Microcomputers.

2.

For a discussion of designing with interrupts, see R. L. Baldridge, "Interrupts Add

Power. Complexity to Microcomputer System Design," EDN, August 5, 1977, pp.
67-73.
-

3.

See Volume 2 of An Introduction to Microcomputers.

4.

See An Introduction to Microcomputers, Volume 2 and 8080A/8085 Assembly

Language Programming.

5.

For further discussion and some real-life examples of designing Z80-based systems
with Interrupts, see pp. 5-24 through 5-37 of Z80 Programming for Logic Design
and the following:
Baldridge, R. L..""lnterrupts Add Power, Complexity to Microcomputer System
Design," EDN. August 5, 1977. pp. 67-73.
Pond, R. M.. "Let Microprocessors Communicate:' Electronic Design, Noyember
1977, pp. 88-90.
--

8.

Shima. M.. and R. Blacksher. "Improved Microprocessor Interrupt Capability:'

Electronic Design, April 26. 1978. pp. 96-100.
Weller. W. J.. Practical Microcomputer Programming: the Z80. Northern Technology Books, Evanston. III.. 1978.
Winston. A. W., and T. B. Smith, "Use of the Z-80 in Data Collection and Control:'
IECI '78 Proceedings - Industrial Applications of Microprocessors, March 20-22,
1978, pp. 208-214.
The Proceedings of the IEEE's Industrial Electronics and Control Instrumentation
Group's Annual Meeting on "Industrial Applications of Microprocessors" contains
many Interesting articles. Volumes (starting with 1975) are available from IEEE Service
Center. CP Department. 445 Hoes Lane. Piscataway, NJ 08854.

12-32

Chapter 13
PROBLEM DEFINITION AND
PROGRAM DESIGN
THE TASKS OF SOFTWARE DEVELOPMENT
In the previous chapters. we have concentrated on the writing of short programs in assembly language. While this is an Important tOpiC. it IS only a small part of software
development. Although writing assembly language programs IS a major task for the
beginner. it soon becomes Simple. By now. you should be familiar with standard
methods for programming In assembly language on the l80 microprocessor. The next
four chapters will describe how to formulate tasks as programs and how to combine short programs to form a working system.

Software development consists of many stages. Figure

13-1 is a flowchart of the software development process. Its
stages are:

STAGES OF
SOFTWARE
DEVELOPMENT

Problem definition
Program design
Coding
Debugging
Testing
Documentation
Maintenance and redeSign
Each of these stages IS Important in the construction of a working system. Note that
coding. the writing of programs in a form that the computer understands. is only one of
seven stages.
In fact. coding is usually the easiest stage to define and perRELATIVE
IMPORTANCE
form. The rules for writing computer programs are easy to learn.
They vary somewhat from computer to computer. but the basIc
OF CODING
techniques remain the same. Few software prOjects run into trouble because of coding: indeed. coding IS not the most time-consuming part of software
development. Experts estimate that a programmer can write one to ten fully debugged
and documented statements per day. Clearly. the mere coding of one to ten statements
is hardly a full day's effort. On most software prolects. coding occupies less than 25% of
the programmer's time.

Measuring progress in the other stages is difficult. You can say

that half of the program has been written. but you can hardly say
that half of the errors have been removed or half of the problem
has been defined. Timetables for such stages as program design.
debugging. and testing are difficult to produce. Many days or weeks of effort may result
in no clear progress. Furthermore. an Incomplete job In one stage may result In tremendous problems later. For example. poor problem definition or program design can make
debugging and testing very difficult. Time saved in one stage may be spent many times
over In later stages.

13-1

Problem definition

Program design

Debugging

Testing

No

No
Documentation

Maintenence end
nedesign

Figure 13-1. Flowchart of Software Develooment

13-2

DEFINITION OF THE STAGES

Problem definition Is the formuletlon of the task in terms of
the requirements that it places on the computer. For example.
what is necessary to make a computer control a tool. run a series
of electrical tests. or handle communications between a central controller and a remote
instrument? Problem definition requires that you determine the forms and rates of Inputs and outputs. the amount and speed of processing that IS needed. and the types of
possible errors and their handling. Problem definition takes a vague idea of building a
computer-controlled system and defines the tasks and requirements for the computer.
Program design is the outline of the computer program which
will perform the tasks that have been defined. In the design
stage. the tasks are described in a way that can easily be converted into a program. Among the useful techniques in this stage are flowcharting.
structured programming. modular programming. and top-down design.
Coding is the writing of the program in a form that the computer can either directly understand or translate. The form may
be machine language. assembly language. or a high-level language.

Ir"'C-O-D-I-N-G""1

Debugging. also called program verification. is making the proDEBUGGING

gram do what the design specified that it would do. In this
stage. you use such tools as breakpoints. traces. simulators. logic analyzers. and in-circuit emulators. The end of the debugging stage is hard to define. since you never know
when you have found the last error.
Testing. also referred to as program validation. is ensuring that
ITESTINGI
the program performs the overall system tasks correctly. The
designer uses simulators. exercisers. and various statistical techniques to get some
measure of the program's performance.

Documentation is the description of the program in the DOCUMENTATION

proper form for users and maintenance personnel. Documentation also allows the designer to develop a program library so that subsequent
tasks will be far simpler. Flowcharts. comments. memory maps. and library forms are
some of the tools used in documentation.

Maintenance and redesign are the servicing. improvement.

MAINTENANCE
AND
and extension of the program. Clearly. the designer must be
ready to handle field problems in computer-based equipment.
REDESIGN
Special diagnostic modes or programs and other maintenance
tools may be reqUired. Upgrading or extension of the program may be necessary to
meet new requirements or handle new tasks.
The rest of this chapter will consider only the problem definition and program
design stages. Chapter 14 will discuss debugging and testing. and Chapter 15 will discuss documentation. extension. and redesign. We will bring all the stages together In
some simple systems examples In Chapter 16.

PROBLEM DEFINITION
Typical microprocessor tasks require a lot of definition. For example. what must a program do to control a scale. a cash register. or a signal generator? Clearly. we have a
long way to go Just to define the tasks involved.

13-3

DEFINING THE INPUTS

How do we start the definition? The obvious place to begin is with the inputs. We
should begin by listing all the inputs that the computer may receive in this application.
Examples of inputs are:
Data blocks from transmission lines
Status words from peripherals
Data from AID converters
Then. we may ask the following questions about each input:
1)

What is ItS form; I.e., what signals will the computer actually
receive?

2)

When is the Input available and how does the processor know It is available? Does
the processor have to request the Input with a strobe signal? Does the input proVide Its own clock?
How long is It available?

3)
4)
5)
6)

How often does it change. and how does the processor know that it has changed?
Does the Input consist of a sequence or block of data? Is the order Important?
What should be done if the data contains errors? These mav include transmission
errors. incorrect data. sequencing errors. extra data. etc.

7)

Is the input related to other Inputs or outputs?

DEFINING THE OUTPUTS

The next step to define IS the output. We must list all the outputs that the computer
must produce. Examples of outputs Include:
Data blocks to transmission lines
Control words to peripherals
Data to 01A converters
Then. we may ask the following questions about each output: .
1)
2)
3)

What IS ItS form; I.e.. what signals must the computer produce?
When must it be available. and how does the peripheral know It IS available?
How long must it be available?

4)
5)

How often must It change. and how does the peripheral know that it has changed?
Is there a sequence of outputs? Is the order Important?

6)

What should be done to avoid transmission errors or to sense and recover from peripheral failures?

7)

How IS the output related to other inputs and outputs?

13-4

PROCESSING SECTION
Between the reading of input data and the sending of output results IS the processing
section. Here we must determine exactly how the computer must process the input data. The questions are:
1)
2)
3)
4)
5)
6)
7)

What IS the basIc procedure (algorithm) for transforming Input

data Into output results?
What time constraints exist? These may Include data rates.
delay times. the time constants of input and output devices. etc.
What memory constraints eXist? Do we have limits on the amount of program
memory or data memory. or on the size of buffers?
What standard programs or tables must be used? What are their requirements?
What special cases exist and how should the program handle them?
How accurate must the results be?
How should the program handle processing errors or special conditions such as
overflow. underflow. or loss of significance?

ERROR HANDLING
An important factqr in many applications is the handling of errors. Clearly. the
deSigner must make provisions for recovering from common errors and for diagnosing
malfunctions. Among the questions that the designer must ask at the definition
stage are:
1)
2)

3)
4)
5)
6)

What errors could occur?

ERROR
Which errors are most likely? If a person operates the CONSIDERATIONS
system. human error IS the most common. Following
human errors. communications or transmiSSion errors are more common than
mechanicaL electricaL mathematical. or processor errors.
Which errors will not be Immediatelv obVIOUS to the system? A special problem IS
the occurrence of errors that the system or operator may not recognize as Incorrect.
How can the system recover from errors with a minimum loss of time and data and
yet be aware that an error has occurred?
Which errors or malfunctions cause the same system behavior? How can these errors or malfunctions be distinguished for diagnostic purposes?
Which errors involve special system procedures? For example. do panty errors reqUire retransmission of data?

Another question is: How can the field techniCian systematically find the source of
malfunctions Without being an expert? Built-In test programs. special diagnostics. or
signature analYSIS can help.1

13-5

HUMAN FACTORS
Many microprocessor-based systems Involve human interaction.
Human factors must be considered throughout the development process for such systems. Among the questions that the
designer must ask are:
1)
2)
3)

4)

OPERATOR
INTERACTION

What input procedures are most natural for the human operator?
Can the operator easily determine how to begin. continue and end the input
operations?
How IS the operator Informed of procedural errors and equipment malfunctions?
What errors is the operator most likely to make?

6)

How does the operator know that data has been entered correctly?
Are displays In a form that the operator can easily read and understand?

7)

Is the response of the system adequate for the operator?

8)

Is the system easy for the operator to use?

Are there guiding features for an inexperienced operator?
Are there shortcuts and reasonable options for the experienced operator?
Can the operator always determine or reset the state of the system after interruptions or distractions?

5)

9)

10)

11)

Building a system for people to use is difficult. The microprocessor can make the
system more powerful. more flexible. and more responsive. However. the designer still
must add the human touches that can greatly increase the usefulness and attractiveness of the system and the productivity of the human operator. 2

EXAMPLES
Response to a Switch
Figure 13-2 shows a simple system in which the input is from
a single SPST switch and the output is to a single LED display.
In response to a switch closure. the processor turns the display on for one second. ThiS system should be easy to define.

DEFINING
SWITCH AND
LIGHT
SYSTEM

Let us first examine the input and answer each of the questions
previously presented:

SWITCH AND
LIGHT INPUT

1)

The Input IS a single bit. which may be either '0' (switch

closed) or '1' (switch open).

2)

The input IS always available and need not be requested.

3)

The input IS available for at least several milliseconds after the closure.

4)

The input will seldom change more than once every few seconds. The processor
has to handle only the bounce In the SWitch. The processor must monitor the
SWitch to determine when it is closed.
There is no sequence of Inputs.

5)
6)

7)

The obVIOUS Input errors are switch failure. failure In the input CirCUitry, and the
operator attempting to close the SWitch again before a sufficient amount of time
has elapsed. We will discuss the handling of these errors later.
The Input does not depend on any other Inputs or outputs.

13-6

+5V

Input

Port

+5V=--=!-

cpu
Output
Port

The switch IOput IS a . r jf the sWitch is open, .O' if the

sWItch is closed. The CPU applies the output to the

cathode of the LED: a '0' lights the display.

Figure 13-2. The Switch and light System

The next requirement in defining the system is to examine the
output. The answers to our questions are:

SWITCH
AND LIGHT
OUTPUTS

1)

The output IS a single bit which is '0' to turn the display on,
'1' to turn it off.

2)

There are no time constraints on the output. The peripheral does not need to be informed of the availability of data.
If the displav is an LED, the data need be available for only a few milliseconds at a
pulse rate of about 100 times per second. The observer will see a continuously lit
display.
The data must change (go off) after one second.
There is no sequence of outputs.
The possible output errors are display failure and failure In the output circuitry.
The output depends only on the switch input and time.

3)

4)
5)
6)
7)

The processing section is extremely simple. As soon as the switch input becomes
a logic '0', the CPU turns the light on (a logic '0') for one second. No time or memory constraints exist.
Let us now look at the possible errors and malfunctions. These
are:

Another sWitch closure before one second has elapsed

SWitch failure
Display failure
Computer failure

SWITCH AND
LIGHT ERROR
HANDLING

Surely the first error IS the most likely. The simplest solution is for the processor to Ignore switch closures until one second has elapsed. This bnef unresponsive penod will
hardly be noticeable to the human operator. Furthermore, ignoring the switch dunng
this period means that no debouncing cirCUitry or software IS necessary, since the
system will not react to the bounce anyway.

13-7

Clearly. the last three failures can produce unpredictable results. The display may stay
on. stay off. or change state randomly. Some possible ways to Isolate the failures would
be:
Lamp-test hardware to check the display; i.e.. a button that turns the light on
independently of the processor
A direct connection to the sWitch to check ItS operation
A diagnostic program that exercises the input and output circuits
If both the display and sWitch are working. the computer is at fault. A field technician
with proper equipment can determine the cause of the failure.

A Switch-Based Memory loader

Figure 13-3 shows a system that allows the user to enter DEFINING A
data into any memory location in a microcomputer. One in- SWITCH-BASED
put port. DPORT, reads data from eight toggle switches. MEMORY LOADER
The other input port. CPORT, is used to read control information. There are three momentary switches: High Address, Low Address and
Data. The output is the value of the last completed entry from the data switches;
eight LEDs are used for the display.
The system will also. of course, require various resistors. buffers. and drivers.

We shall first examine the inputs. The characteristics of the sWitches are the same as
in the previous example; however. here there IS a distinct sequence of inputs, as
follows:
1)
2)
3)
4)
5)
6)

The operator must set the data sWitches according to the eight most significant
bits of an address, then
press the High Address button. The high address bits will &ppear on the lights, and
the program will interpret the data as the high byte of the address.
Then the operator must set the data switches with the value of the least significant
byte of the address and
press the Low Address button. The low address bits will appear on the lights, and
the program will consider the data to be the low byte of the address.
Finally, the operator must set the desired data into the data sWitches and
press the Data button. The display will now show the data. and the program stores
the data in memory at the previously entered address.

The operator may repeat the process to enter an entire program. Clearly, even in this
simplified situation, we will have many possible sequences to consider. How do we
cope with erroneous sequences and make the system easy to use?

Output is no problem. After each input, the program sends to the displays the
complement (since the displays are active-Iowl of the input bits. The output data
remainS the same until the next input operation.
The processing section remains quite simple. There are no time or memory constraints. The program can debounce the switches by waiting for a few milliseconds. and
must provide complemented data to the displays.

13-8

Input
Port

DPORT

Data
Bus

High Address

Input

CPU

Low Address

Port

CPORT
Data

+5V

Output

Port

Figure 13-3. The Switch-Based Memory Loader

13-9

The most likely errors are operator mistakes. These include:

Incorrect entries
Incorrect order
Incomplete entries; for example. forgetting the data

MEMORY

LOADER
ERROR
HANDLING

The system must be able to handle these problems in a reasonable

wav. since they are certain to occur In actual operation.
The designer must also consider the effects of equipment failure. Just as before.
the possible difficulties are:
Switch failure
Display failure
Computer failure
In this svstem. however. we must pav more attention to how these failures affect the
system. Acomputer failure will presumably cause very unusual behavior by the system.
and will be easy to detect. A display failure may not be immediately noticeable; here a
Lamp Test feature will allow the operator to check the operation. Note that we would
like to test each LED separately. in order to diagnose the case in which output lines are
shorted togethe~ In addition. the operator may not immediately detect switch failure;
however. the operator should soon notice it and establish which switch is faulty by a
process of elimination.
Let us look at some of the possible operator errors. Typical errors
will be:

OPERATOR
ERROR
CORRECTION
IN MEMORY
LOADER

Erroneous data
Wrong order of entries or sWitches
Trying to go on to the next entry Without completing the current
one
The operator will presumably notice erroneous data as soon as it appears on the displays. What IS a viable recovery procedure for the operator? Some of the options are:
1)

2)

3)

The operator must complete the entry procedure; I.e.. enter Low Address and Data
if the error occurs In the High Address. Clearly. this procedure is wasteful and
would only serve to annoy the operator.
The operator mav restart the entry process by returning to the high address entry
steps. ThiS solution IS useful if the error was in the High Address. but forces the
operator to re-enter earlier data if the error was in the Low Address or Data stage.
The operator may enter any part of the sequence at any time simply by setting the
Data switches With the desired data and pressing the corresponding button. ThiS
procedure allows the operator to make corrections at any point in the sequence.

ThiS type of procedure should always be preferred over one that does not allow Immediate error correction. has a variety of concluding steps. or enters data Into the svstem
without allowing the operator a final check. Any added complication in hardware or
software will be justified in increased operator efficiency. You should always prefer to
let the microcomputer do the tedious work and recognize arbitrary sequences; It never
gets tired and never forgets what was in the operating manual.
A further helpful feature would be status lights that would define the meaning of the
displav. Three status lights. marked "High Address", "Low Address", and "Data".
wou Id let the operator know what had been entered without haVing to remember which
button was pressed. The processor would have to monitor the sequence. but the added
complication in software would simplify the operator's task. Clearly. three separate sets
of displays pius the ability to examine a memory location would be even more helpful to
the operator.

13-10

,---

Kevboard

Input Port

Kevboard Strobe

A-

I Kevboard Data

.....

~
CPU

---.

Dispiay

.....

Output POrUsi

"Y

> Displav

.-

XMIT
Output Port

"--

RCV

Input Port

Data Strobe
..oL

From Central Computer

.....

.
Status Light
Output Port

Peripheral Readv Strobe

To Central Computer

BUSY Displav
READY Displav

Figure 13-4. Block Diagram of a Verification Terminal

We should note that, although we have emphasized human interaction, machine
or system interaction has many of the same characteristics. The microprocessor
should do the work. If complicating the microprocessor's task makes error recovery simple and the causes of failure obvious, the entire system will work better
and be easier to maintain. Note that you should not wait until after the software has
been completed to consider system use and maintenance: instead. yOU should include
these factors in the problem definition stage.

A Verification Terminal
DEFINING A
Figure 13-4 is a block diagram of a simple credit-verification
VERIFICAnON
terminal. One input port derives data from a keyboard (see
Figure 13-5): the other input port accepts verification data
TERMINAL
from a transmission line. One output port sends data to a set of
displays (see Figure 13-6): another sends the credit card number to the central
computer. A third output port turns on one light whenever the terminal is ready to
accept an inquiry, and another light when the operator sends the information. The
"Busy" light turns off when the response returns. Clearly. the Input and output of
data will be more complex than In the previous case. although the processing is still
simple.

13-11

ODD

DDDB

DOOG
D

The digIt keys allow digIt entnes.

CLEAR deletes the entIre entry.

SEND transmIts the entry to the central computer.

Figure 13-5. Verification Terminal Keyboard

BUSY

READY

The display conSists of ten 7-segment displays. whIch may be multiplexed. controlled bv a shift
register, or addressed separateIv. Two additional lights, READY end BUSY, are also present.

Figure 13-6. Verification Terminal Display

13-12

Additional displays may be useful to emphasize the meaning of the response. Many terminals use a green light for "Yes". a red light for "No". and a yellow light for "Consult
Store Manager." Note that these lights will still have to be clearly marked with their
meanings to allow for a color-blind operator.
VERIFICATION
Let us first look at the keyboard input. This is. of course,
TERMINAL
different from the switch input. since the CPU must have some
INPUTS
way of distinguishing new data. We will assume that each key
closure provides a unique hexadecimal code (we can code
each of the 12 keys into one digiti and a strobe. The program will have to recognize the strobe and fetch the hexadecimal number that identifies the key. There IS a
time constraint. since the program cannot miss any data or strobes. The constraint IS
not serious. since keyboard entries will be at least several milliseconds apart.
The transmission input similarly consists of a series of characters. each identified
by a strobe (perhaps from a UARTI. The program will have to recognize each
strobe and fetch the character. The data being sent across the transmission lines
is usually organized into messages. A possible message format is:

Introductory characters, or header

Terminal destination address
Coded yes or no
Ending characters, or trailer
The terminal will check the header. read the destination address, and see if the
message is intended for it. If the message is for the terminal. the terminal accepts the
data. The address could be (and often is) hard-wired into the terminal so that the terminal receives only messages Intended for it. This approach simplifies the software at
the cost of some flexibility.
VERIFICATION
The output is also more complex than in the earlier examples.
TERMINAL
If the displays are multiplexed. the processor must not only
OUTPUTS
send the data to the display port but must also direct the data
to a particular display. We will need either a separate control port
or a counter and decoder to handle this. Note that hardware blanking controls can
blank leading zeros as long as the first digit in a multi-digit number is never zero. Software can also handle this task. Time constraints include the pulse length and frequency
required to produce a continuous display for the operator.
The communications output will consist of a series of characters with a particular
format. The program will also have to consider the time required between characters. A possible format for the output message is:

Header
Terminal address
Credit card number
Trailer
A central communications computer may poll the terminals. checking for data
ready to be sent.
The processing in this system involves many new tasks. such as:

IdentifYing the control keys by number and performing the proper actions
Adding the header, terminal address, and trailer to the outgOing message
Recognizing the header and trailer in the returning message
Checking the Incoming terminal address

13-13

Note that none of the tasks involve any complex arithmetic or any
serious time or memory constraints.

The number of possible errors in this system is, of course,

much larger than in the earlier examples. Let us first consider
the possible operator errors. These include:

VERIFICATION
TERMINAL
ERROR
HANDLING

Entering the credit card number IncorrectIv

Trying to send an incomplete credit card number
Trying to send another number while the central computer is processing one
Clearing non-existent entries
Some of these errors can be easily handled by correctly structuring the program. For example, the program should not accept the Send key until the credit card number has
been completely entered. and It should ignore any additional keYboard entries until the
response comes back from the central computer. Note that the operator will know that
the entry has not been sent. since the Busy light will not go on. The operator will also
know when the keyboard has been locked out (the program is ignoring keyboard entries!. since entries will not appear on the display and the Ready light will be off.
Incorrect entries are an obvious problem. If the operator recogCORRECTING
nizes an error. he can use the Clear key to make corrections. The
KEYBOARD
operator would probably find It more convenient to have two Clear
ERRORS
keys, one that cleared the most recent key and one that cleared
the entire entry. This would allow both for the situation in which the operator recognizes the error immediately and for the situation in which the operator recognizes the
error late in the procedure. The operator should be able to correct errors immediately
and have to repeat as few keys as possible. The operator will. however, make a certain
number of errors without recognizing them. Most credit card numbers include a selfchecking digit: the terminal could check the number before permitting it to be sent to
the central computer. This step wou Id save the central computer from wasting precious
processing time checking the number.
This requires. however. that the terminal have some way of informing the operator of
the error. perhaps by flashing one of the displays or by providing some other special indicator that the operator is sure to notice.
Still another problem is how the operator knows that an entry has been lost or processed Incorrectly. Some terminals simply unlock after a maximum time delay. The
operator notes that the Busy light has gone off Without an answer being received. The
operator is then expected to try the entry again. After one or two retries, the operator
should report the failure to supervisory personnel.

Many equipment failures are also possible. Besides the displays, keyboard, and
processor, there now exist the problems of communications errors or failures and
central computer failures.

13-14

The data transmission will probably have to Include error checking and correcting procedures. Some possibilities are:
1)

Parity provides an error detection facility but no correction

mechanism. The receiver will need some way of requesting retransmiSSion. and the sender will have to save a copy
of the data until proper reception IS acknowledged. Panty
IS. however. very simple to Implement.

2)

Short messages may use more elaborate schemes. For example. the yes/no
response to the terminal could be coded so as to provide error detection and correction capability.

3)

An acknowledgement and a limited number of retries could trigger an indicator

that would inform the operator of a communications failure (inability to transfer a
message without errors) or central computer failure (no response' at all to the
message within a certain period of time!. Such a scheme. along with the Lamp
Test. would allow simple failure diagnosis.

CORRECTING
TRANSMISSION
/ ERRORS

A communications or central computer failure indicator should also "unlock" the terminal. Le.. allow it to accept another entry. ThiS is necessary if the terminal will not accept entries while a verification is in progress. The terminal may also unlock after a certain maximum time delay. Certain entries could be reserved for diagnostics: Le.. certain
credit card numbers could be used to check the internal operation of the terminal and
test the displays.

REVIEW OF PROBLEM DEFINITION

Problem definition is as important a part of software development as it is of any
other engineering task. Note that it does not require any programming or
knowledge of the computer; rather, it is based on an understanding of the system
and sound engineering judgment. Microprocessors can ofter flexibility that the
designer can use to provide a range of features which were not previously available.
Problem definition is independent of any particular computer, computer language,
or development system. It should, however, provide guidelines as to what type or
speed of computer the application will require and what kind of hardware/software trade-ofts the designer can make. The problem definition stage is
in fact independent of whether or not a computer is used at all, although a
knowledge of the capabilities of the computer can help the designer in suggesting
possible implementations of procedures.

13-15

PROGRAM DESIGN
Program design is the stage in which the problem definition is formulated as a program. If the program is small and simple. this stage may involve little more than
the writing of a one-page flowchart. If the program is larger or more complex. the
designer should consider more elaborate methods
We will discuss flowcharting. modular programming. structured programming. and
top-down design. We will try to indicate the reasoning behind these methods. and
their advantages and disadvantages. We will not, however. advocate any particular
method since there is no evidence that one method is always superior to all others. You
should remember that the goal is to produce a good working system. not to follow
religiously the tenets of one methodology or another.
All the methodologies do, however. have some obvious principles in common. Many of these are the same principles that apply
to any kind of design. such as:

1)
2)

3)
4)
5)
6)
7)

8)

BASIC
PRINCIPLES
OF PROGRAM
DESIGN

Proceed In small steps. Do not try to do too much at one

time.
Divide large lobs Into small. logicallY separate tasks. Make the sub-tasks as independent of one another as possible, so that they can be tested separately and so
that changes can be made in one without affecting the others.
Keep the flow of control as simple as possible so as to make it easier to find errors.
Use pictOrial or graphic descriptions as much as possible. They are easier to
Visualize than word descriptions. This is the great advantage of flowcharts.
Emphasize clarity and simplicity at first. You can Improve performance (if necessary) once the system IS working.
Proceed in a thorough and systematic manner. Use checklists and standard procedures.
Do not tempt fate. Either do not use methods that you are not sure of. or use them
very carefully. Watch for Situations that might cause confUSion. and clarify them
as soon as possible.
Keep in mind that the system must be debugged. tested and maintained. Plan for
these later stages.

9)

Use simple and consistent terminology and methods. Repetitiveness is no fault In

program design, nor is complexity a virtue.

10)

Have your design completely formulated before you start coding. Resist the
temptation to start writing down instructions: it makes no more sense than making parts lists or laYing out circuit boards before you know exactly what will be In
the system.

11)

Be particu larly carefu I of factors that may change. Make the implementation of
likely changes as simple as possible.

13-16

Input/Output

o
<:>
oo
1
(

Processing operation
(Arithmetic. LogIC. Data Movement)

Decision logiC

Subroutine

Connector pOint

Connector arrows

Terminal pomt

(Beginning or EndingJ

Figure 13-7 Standard Flowchart Symbols

FLOWCHARTING
Flowcharting is certainly the best-known of all program design methods. Programming
textbooks describe how programmers first write complete flowcharts and then start
writing the actual program. In fact. few programmers have ever worked this way, and
flowcharting has often been more of a loke or a nuisance to programmers than a design
method. We will try to describe both the advantages and disadvantages of flowcharts,
and show the place of this technique in program design.

The basic advantage of the flowchart is that it is a pictorial

ADVANTAGES OF
representation. People find such representations much more
FLOWCHARTING
meaningfuL than written descriptions. The designer can visualIze the whole system and see the relationships of the various parts. Logical errors and
inconsistencies often stand out instead of being hidden In a printed page. At its best.
the flowchart is a picture of the entire system.

13-17

Some of the more specific advantages of flowcharts are:

1)
2)
3)
4)
5)
6)

Standard symbols exist (see Figure 13-7) so that flowcharting forms are widely
recognized.
Flowcharts can be understood by someone without a programming background.
Flowcharts can be used to divide the entire project mto sub-tasks. The flowchart
can then be exammed to measure overall progress.
Flowcharts show the sequence of operations and can therefore aid m locating the
sou rce of errors.
Flowcharting is widely used m other areas besides programming.
There are many tools available to aid In flowcharting. including programmer's
templates and automated drawmg packages.

These advantages are all important. There is no question that

flowcharting will continue to be widely used. But we should
note some of the disadvantages of flowcharting as a program design method. e.g.:
1)

2)
3)

4)
5)
6)

DISADVANTAGES
OF
FLOWCHARTING

Flowcharts are difficult to deSign, draw, or change in all except the simplest situations.
There is no easy way to debug or test a flowchart.
Flowcharts tend to become cluttered. DeSigners find it difficult to balance between
the amount of detail needed to make the flowchart useful and the amount that
makes the flowchart little better than a program listing.
Flowcharts show only the program organization. They do not show the organization
of the data or the structure of the input/output modules.
Flowcharts do not help with hardware or tlmmg problems or give hmts as to where
these problems might occur.
Flowcharts allow for highly unstructured design. Lines and arrows backtrackmg
and looping all over the chart are the antithesis of good structured design princIples.

Thus, flowcharting is a helpful technique that you should not try to extend too far.
Flowcharts are useful as program documentation. since they have standard forms
and are comprehensible to non-programmers. As a deSign tool. however. flowcharts
cannot provide much more than a starting outline; the programmer cannot debug a
detailed flowchart and the flowchart is often more difficult to design than the program
itself.

13-18

eXAMPLES
Response to a Switch
This simple task, in which a single switch turns on a light
for one second, is easy to flowchart. In fact. such tasks are
typical examples for flowcharting books, although they form a
small part of most systems. The data structure here IS so simple
that It can be safely Ignored.

FLOWCHARTING
SWITCH AND
LIGHT SYSTEM

Figure 13-8 is the flowchart. There is little difficulty In deciding on the amount of
detail required. The flowchart gives a straightforward picture of the procedure, which
anyone could understand.
Note that the most useful flowcharts may Ignore program vanables and ask questions
directly. Of course, compromises are often necessary here. Two versions of the
flowchart are sometimes helpful- one general version in layman's language,
which will be useful to non-programmers, and one programmer's version in terms
of the program variables, which will be useful to other programmers.

A third type of flowchart, a data flowchart, may also be ~~--DATA

helpful. This flowchart serves as a cross-reference for the other
FLOWCHARTS
flowcharts, since It shows how the program handles a particular
type of data. Ordinary flowcharts show how the program proceeds, handling different
types of data at different POints. Data flowcharts, on the other hand, show how particular types of data move through the system, passing from one part of the program to
another. Such flowcharts are very useful in debugging and maintenance, since errors
most often show up as a particular type of data being handled incorrectly.

13-19

Turn light on

Turn light off

End

Figure 13-8. Flowchart of One-Second Response to a SWitch

The Switch-Based Memory loader

This system (see Figure 13-3) IS considerably more complex
than the previous example. and involves many more deCISions.
The flowchart (see Figure 13-91 is more difficult to write
and not as straightforward as the previous example. In this
example. we face the problem that there IS no way to debug or
test the flowchart.

FLOWCHARTING
THE
SWITCH-BASED
MEMORY LOADER

The flowchart In Figure 13-9 Includes the improvements we suggested as part of the
problem definition. Clearly. this flowchart is beginning to get cluttered and lose its
advantages over a written description. Adding other features that define the meaning of the entry with status lights and allow the operator to check entries after completion would make the flowchart even more complex. Writing the complete flowchart
from scratch could quicklV become a formidable task. However. once the program has
been written. the flowchart IS useful as documentation.

13-20

High bvte of
Address =
Switches

Ught. = Switches

Walt

Low bvte of
Address =
SWitches

lights

= SvvitchE:s

debounce

Wait
debounce

time

time

Data =SWltches

Ughts = Switches

Store Data
at Address

Walt
debounce

time
No

No

Walt
debounce

time

Figure 13-9. Flowchart of Switch-Based Memory Loader

13-21

Oasr Entry Array

Key Pointer = Start

of Entrv Array
Key Coonter

=0

Kev= Keyboard

Input Oata
(Key PoInter)

Kev Pointer

= Kev

Key -Point... + 1
Key Counter =
Key Counter + 1

Figure 13-10. Flowchart of Keyboard Entry Process

The Credit-Verification Terminal

In this application (see Figures 13-4 through 13-6), the
flowchart will be even more complex than In the switch-based
memory loader case. Here, the best idea is to flowchart sections separately so that the flowcharts remain manageable,
However, the presence of data structures (as In the multi-digit
display and the messages) will make the gap between
flowchart and program much wider.

FLOWCHARTING
THE CREDIT
VERIFICATION
FLOWCHARTING
SECTIONS

Let us look at some of the sections. Figure 13-10 shows the keyboard entry process
for the digit keys. The program must fetch the data after each strobe and place the
digit into the display array if there is room for It. If there are already ten digits In the array, the program Simply ignores the entry.
The actual program will have to handle the displays at the same time. Note that either
software or hardware must de-activate the keyboard strobe after the processor reads a
digit.

13-22

Kev = Keyboard
Input Data

Figure 13-11. Flowchart of Keyboard Entry Process with Send Key

Figure 13-11 adds the Send key. This key. of course. IS optional. The terminal could
just send the data as soon as the operator enters a complete number. However. that
procedure would not give the operator a chance to check the entire entry. The
flowchart with the Send key is more complex because there are two alternatives.

1)

If the operator has not entered ten digits. the program must Ignore the Send key
and place anv other key into the entrv.

2)

If the operator has entered ten digits. the program must respond to the Send kev by
transferring control to the Send routine. and Ignore all other keys.

Note that the flowchart has become much more difficult to organize and to follow.
There is also no obvious wav to check the flowchart.

13-23

Clear Display Array

Kev PoIntar = Start
of Display Array
Kev Counter 0

Kay = Keyboard
Input Data

(Kev POinter)

= Key

Key Pointer =

Key PcMnter + 1

Figure 13-12. Flowchart of Keyboard Entry Process with Function Keys

Figure 13-12 shows the flowchart of the keyboard entry process with all the function keys. In this example. the flow of control IS not simple. Clearly. some written
desCriptIOn IS necessary. The organization and layout of complex flowcharts requires
careful planning. We have followed the process of adding features to the flowchart one
at a time. but this still results In a large amount of redrawing. Again we should remember that throughout the keyboard entry process. the program must also refresh the displays if they are multiplexed and not controlled by shift registers or other hardware.

13-24

AJ---......
Header flag = 0
Parity Error flag = 0
Address Match f1ag=O
Address Pointer =Sta

of terminal address
Address Counter = 0

Nmess =0

Panty Error
flag = 1

Messg INmessi =
Character
Nmess = Nmess +

Header flag = 1

Address Counter =
Address Counter + 1
Address POinter =
Address Pointer + 1

Tum off Busv

light

Figure 13-13. Flowchart of Receive Routine

13-25

Figure 13-13 is the flowchart of a receive routine. We assume that the serra I/para lie I
conversion and error checking are done in hardware (e.g.. by a UART). The processor
must:
1)
2)

3)
4)
5)

Look for the header (we assume that It IS a single character).

Read the destination address (we assume that It is three characters long) and see if
the message IS meant for this terminal: I.e.. if the three characters agree with the
terminal address.
Wait for the trailer character.
If the message is meant for the terminal, turn off the Busy light and go to Display
Answer routine.
In the event of any errors. request retransmission by going to RTRAN routine.

This routine Involves a large number of decisions, and the flowchart IS neither simple
nor obvious.
Clearly, we have come a long way from the simple flowchart (Figure 13-8) of the
first example. A complete set of flowcharts for the transaction terminal would be
a major task. It would consist of several interrelated charts with complex logic. and
would require a large amount of effort. Such an effort would be just as difficult as writIng a preliminary program. and not as useful. since you could not check it on the computer.

MODULAR PROGRAMMING
Once programs become large and complex. flowcharting is no longer a satisfactory
deSign tool. However, the problem definitIOn and the flowchart can give you some idea
as to how to diVide the program Into reasonable sub-tasks. The division of the entire
program into sub-tasks or modules is called "modular programming." Clearly. most
of the programs we presented In earlier chapters would typically be modules In a large
system program. The problems that the designer faces in modular programming are
how to divide the program into modules and how to put the modules together.
The advantages of modular programming are obvious:

ADVANTAGES

1)

A single module is easier to write. debug, and test than an

entire program.

OF MODULAR
PROGRAMMING

2)

A module IS likely to be useful in many places and In other

programs. particularly if it IS reasonably general and performs a common task. You
can build up a library of standard modules.
Modular programming allows the programmer to diVide tasks and use previously
written programs.

3)
4)
5)
6)

Changes can be Incorporated into one module rather than into the entire system.
Errors can often be Isolated and then attributed to a Single module.
Modular programming gives an Idea of how much progress has been made and
how much of the work IS left.

13-26

The idea of modular programming is such an obvious one

that its disadvantages are often ignored. These include:

DISADVANTAGES
OF MODULAR
PROGRAMMING

1)

Fitting the modules together can be a major problem. particularly if different people wnte the modules.

2)

Modules require very careful documentation. since they may affect other parts of
the program. such as data structures used by all the modules.
Testing and debugging modules separately IS difficult. since other modules may
produce the data used by the module being debugged and still other modules may
use the results. You may have to write special programs (called "dnvers") Just to
produce sample data and test the programs. These drivers require extra programming effort that adds nothing to the system.
Programs may be very difficult to modulanze. If you modulanze the program poorly.
integration will be very difficult. since almost all errors and changes will Involve
several modules.
Modular programs often require extra time and memory. since the separate
modules may repeat functions.

3)

4)

5)

Therefore. while modular programming is certainly an improvement over trying to wnte

the entire program from scratch. It does have some disadvantages as well.

Important considerations include restricting the amount of information shared by

modules, limiting design decisions that are subject to change to a single module
and restricting the access of one module to another. 3
An obvious problem is that there are no proven,
systematic methods for modularizing programs. We
should mention the following principles: 4
1)
2)
3)

PRINCIPLES OF
MODULARIZATION

Modules that reference common data should be parts of the same overall module.
Two modules In which the first uses or depends on the second. but not the reverse.
should be separate.
A module that IS used by more than one other module should be part of a different
overall module than the others.

4)

Two modules In which the first IS used by many other modules and the second IS
used by only a few other modules should be separate.

5)

Two modules whose frequencies of usage are significantly different should be part
of different modu les.

6)

The structure or organization of related data should be hidden within a single

module.

If you find it very difficult to modularize your program, it is a strong indication that
the problem is poorly defined, and redefinition is called for. Too many special cases.
each requiring special handling. or the use of a large number of variables. each requirIng special processing. are problems that can be most efficiently handled. by redefining
the tasks at hand.

13-27

EXAMPLES
Response to a Switch
This simple program can be divided into two modules:
Module 1 waits for the switch to be turned on and turns
the light on in response.

MODULARIZING
THE SWITCH
AND LIGHT
SYSTEM

Module 2 provides the one-second delay.

Module 1 is likely to be specific to the system. since It will depend on how the sWitch
and light are attached. Module 2 will be generally useful. since many tasks require
delays. Clearly. it would be advantageous to have a standard delay module that could
provide delays of varying lengths. The module will require careful documentation so
that vou will know how to specifv the length of the delay. how to call the module. and
what registers and memory locations the module affects.
A general version of Module 1 would be far less useful. since It would have to deal with
different types and connections of switches and lights.
You would probably find it simpler to write a module for a particular configuration of
sWitches and lights rather than try to use a standard routine. Note the difference between thiS situation and Module 2.

The Switch-Based Memory loader

The switch-based memory loader is difficult to modularize,
since all the programming tasks depend on the hardware
configuration and the tasks are so simple that modules
hardly seem worthwhile. The flowchart In Figure 13-9 suggests that one module might be the one that walts for the
operator to press one of the three pushbuttons.

MODULARIZING
THE
SWITCH- BASED
MEMORY LOADER

Some other modules might be:

A delay module that provides the delay required to debounce the sWitches
A sWitch and display module that reads the data from the switches and sends it to
the displays
A Lamp Test module
Highly system-dependent modules such as the last two are unlikely to be generally
useful. ThiS example IS not one In which modular programming offers great advantages.

The Verification Terminal

The verification terminal, on the other hand. lends Itself very
well to modular programming. The entire system can easily be
divided into three main modules:
Keyboard and display module
Data transmission module
Data reception module

MODULARIZING
THE
VERIFICATION
TERMINAL

A general keyboard and display module could handle many keyboard- and displaybased systems. The sub-modules would perform such tasks as:
Recognizing a new keyboard entry and fetching the data
Clearing the array in response to a Clear key
Entering digits into storage
Looking for the terminator or Send key
0 isplaYlng the dig its

13-28

Although the key Interpretations and the number of digits will vary. the basic entry.
data storage. and data display processes will be the same for many programs. Such
function keys as Clear would also be standard. Clearly. the designer must consider
which modules will be useful in other applications, and pay careful attention to
those modules.
The data transmission module could also be divided Into such sub-modules as:
1)

Adding the header character.

2)
3)
4)
5)

Transmitting characters as the output line can handle them.

Generating delay times between bits or characters.
Adding the trailer character.
Checking for transmission failures: I.e.. no acknowledgement or Inability to
transmit without errors.

The data reception module could Include sub-modules which:

1)
2)
3)
4)
5)

Look for the header character.

Check the message destination address against the terminal address.
Store and interpret the message.
Look for the trailer character.
Generate bit or character delays.

INFORMATION
Note here how Important it IS that each design decision (such as
the bit rate. message format. or error-checking procedure) be ImHIDING
PRINCIPLE
plemented In only one module. A change I.n any of these decisions
will then require changes only to that Single module. The other
modules should be written so that they are totally unaware of the values chosen or the
methods used in the Implementing module. An important concept here is the "information-hiding principle:'5 whereby modules share only information that is absolutely essential to getting the task done. Other information is hidden within a
single module.
An Important use of thiS principle is In error handling. Whenever a module detects a
lethal error. It should not undertake recovery procedures. Instead. it should pass the error status back up to the calling module and allow It to make the decIsion of how to
recover from the error. The reason for thiS IS that the lower level procedure often does
not have enough Information to adequately decide what recovery procedures are
necessary. For example. suppose we have a module that accepts numeric Input from a
user. ThiS module terminates normally when the user enters a string of numeric digits
terminated by a carnage return. Entry of any non-numeric characters causes the
module to Immediately terminate abnormally. Since the module does not know In what
context It is being used (i.e.. IS It part of an assembler. an Interactive editor. or a file
management system?) It cannot make a valid deCISion of what action to take when encountering an Invalid character. If a single error recovery method was designed Into the
module. it would lose ItS generality and become specifiC to those situations that employ
this error recovery technique.

13-29

REVIEW OF MODULAR PROGRAMMING

Modular programming can be very helpful if you abide by
the following rules:
1)

2)

3)

4)

5)

RULES FOR

MODULAR
PROGRAMMING

Use modules of 20 to 50 lines. Shorter modules are

usually a waste of time, while longer modules are seldom
general and may be difficult to Integrate.
Try to make modules reasonably general. Differentiate between common
features like ASCII code or asynchronous transmiSSion formats. which will be the
same for many applications and key identifications. and number of displays or
number of characters In a message, which are likely to be unique to a particular application. Make the changing of the latter parameters simple. Malor changes like
different character codes should be handled by separate modules.
Take extra time on modules like delays. display handlers, keyboard handlers. etc.
that will be useful in other projects or in many different places in the present
program.
Try to keep modules as distinct and logically separate as possible. Restrict the
flow of information between modules and Implement each design decision in a
single module.
Do not try to modularize simple tasks where rewriting the entire task may be
easier than assembling or modifYing the module.

STRUCTURED PROGRAMMING
How do you keep modules distinct and stop them from interacting? How do you
write a program that has a clear sequence of operations so that you can isolate
and correct errors? One answer is to use the methods known as "structured programming" whereby each part of the program consists of elements from a limited
set of structures and each structure has a single entry and a single exit.
Figure 13-14 shows a flowchart of an unstructured program. If an error occurs in
Module B. we have five possible sources for that error. Not only must we check each sequence, but we also have to make sure that any changes made to correct the error do
not affect any of the other sequences. The usual result IS that debugging becomes like
wrestling an octopus. Every time you think the situation IS under control. there IS
another loose tentacle somewhere.

13-30

Figure 13-14. Flowchart of an Unstructured Program

The solution is to establish a clear sequence of operations so
that you can isolate errors. Such a sequence uses single-entry,
single-exit modules. The basic modules that are needed are:
1)

An ordinary sequence; I.e.. a linear structure In which

statements or structures are executed consecutively. In
the sequence:

BASIC
STRUCTURES
OF
STRUCTURED

PROGRAMMING

S1
S2
S3
the computer executes S1 first. S2 second, and S3 third. S1. S2. and S3 may be
single instructions or entire programs.
2)

3)

A conditional structure.
The common one is "if C then S1 else S2." where C is a condition and S1 and S2
are statements or sequences of statements. The computer executes S1 if C IS true,
and S2 if C is false. Figure 13-15 shows the logic of this structure. Note that the
structure has a single entry and a single eXit: there IS no way to enter or leave S1 or
S2 other than through the structure.
A loop structure.
The common loop structure IS "while C do S," where C IS a condition and S IS a
statement or sequence of statements. The computer checks C and executes S if C
IS true. This structure (see Figure 13-16) also has a Single entry and a Single eXIt.
Note that the computer will not execute S at all if C IS onglnally false, since the
value of C IS checked before S is executed.

13-31

S2

Sl

Figure 13-15. Flowchart of the If-Then-Else Structure

Figure 13-16. Flowchart of the Do-While Structure

In most structured programming languages. an alternative looping construct is provided. This construct is known as the do-until clause, Its basic structure is "do S until
C", where C is a condition and S is a statement or sequence of statements. It is similar
to the do-while construct except that the test of the looping condition C is performed at
the end of the loop. This has the effect of guaranteeing that the loop is alwavs executed
at least once. This IS illustrated by the flowchart in Figure 13-17 The common Indexcontrolled or DO loop can be Implemented as a special case of either of these two baSIC
looping constructs.

13-32

Start

Figure 13-17 Flowchart of the Do-Until Structure

4)

A case structure.
Although not a primitive structure like sequential. if-then-else. and do-while. the
case structure is so commonly used that we Include it here as an adjunct to the
basIc structure descriptions. The case structure is "case I of SO. S1. .Sn". where I
IS an index and SO. S1.
.Sn are statements or sequences of statements. If I IS
equal to zero then statement SO is executed: if I is equal to 1 then statement S1 is
executed. etc. Only one of the n statements IS executed. After ItS execution. control
passes to the next sequential statement following the case statement group. If I IS
greater than n (j.e.. the number of statements In the case statement). then none of
the statements in the case statement IS executed. and control IS passed directly to
the next sequential statement following the case statement. This is illustrated by
the flowchart In Figure 13-18.

Note the following features of structured programming:

11

21
31

Only the three basic structures, and possibly a small number of auxiliary
structures, are permitted.
Structures may be nested to any level of complexity so that any program can,
in turn, contain any of the structures.
Each structure has a single entry and a single exit.

Some examples of the conditional structure illustrated in

Figure 13-15 are:
1)

S2 Included:
if X ;;:::. 0 then NPOS = NPOS
else NNEG = NNEG + 1

EXAMPLES
OF
STRUCTURES

+1

Both S1 and S2 are single statements.

2)

S2 omitted:
if X;bO then Y = 1/X

Here no action is taken if C (X ;b0) is false. S2 and "else" can be omitted in this case.

13-33

No

so

S1

Sn

' - - - - - - I I......---~....---- - - - - -...- - ~

Figure 13-18. Flowchart of the Case Structure

Some examples of the loop structure illustrated in Figure 13-16 are:

1)

Form the sum of integers from 1 to N.

1=0
SUM =0
do while 1< N

1=1+1
SUM = SUM
end

+I

The computer executes the loop as long as f

while" is not executed at all.
2)

< N. If N = O. the program within the "do-

Count characters in an array SENTENCE until you find an ASCII period.

NCHAR =0
do while SENTENCE (NCHAR) ~ PERIOD
NCHAR = NCHAR + 1
end

The computer executes the loop as long as the character in SENTENCE is not an ASCII
period. The count is zero if the first character is a period.

13-34

The advantages of structured programming are:

ADVANTAGES OF
STRUCTURED
PROGRAMMING

11

The sequence of operations is simple to trace. This allows

you to test and debug easily.

21

The number of structures IS limited and the terminology is

standardized.
The structures can easily be made into modules.
Theoreticians have proved that the given set of structures is complete: that is. all
programs can be written In terms of the three structures.
The structured version of a program is partly self-documenting and fairlY easy to
read.
Structured programs are easy to describe with program outlines.
Structured programming has been shown in practice to increase programmer productivity.

31
4)
5)
6)
71

Structured programming basically forces much more discipline on the programmer

than does modular programming. The result is more systematic and betterorganized programs.
The disadvantages of structured programming are:

11

DISADVANTAGES
OF
Only a few high-level languages (e.g.. PLIM. PASCAL) will
STRUCTURED
directly accept the structures. The programmer therefore
PROGRAMMING
has to go through an extra translation stage to convert the
structures to assembly language code. The structured version of the program. however. is often useful as documentation.

2)

Structured programs often execute more slowlY and use more memory than
unstructured programs.

3)

Limiting the structures to the four basic forms makes some tasks very awkward to
perform. The completeness of the structures only means that all programs can be
Implemented with them: It does not mean that a given program can be Implemented efficiently or conveniently.

4)

The standard structures are often quite confUSing. e.g.. nested "if-then-else" structures may be very difficult to read. since there may be no clear indication of where
the Inner structures end. A series of misted "do-while" loops can also be difficult to
read.

5)

Structured programs consider only the sequence of program operations. not the
flow of data. Therefore. the structures may handle data awkwardly.
Few programmers are accustomed to structured programming. Many find the standard structures awkward and restrictive.

6)

We are neither advocating nor discouraging the use of structured programming. It

is one way of systematizing program design. In general. structured programming
is most useful in the following situations:

Larger programs. perhaps exceeding 1000 instructions

Applications In which memory usage IS not CriticaL
Low-volume applications where software development costs.
particularly testing and debugging. are important factors.
Applications involving string manipulation. process control.
or other algonthms rather than Simple bit manipulations.

13-35

WHEN TO USE
STRUCTURED
PROGRAMMING

In the future. we expect the cost of memory to decrease. the average size of
microprocessor programs to increase. and the cost of software development to
increase. Therefore. methods like structured programming. which decrease software development costs for larger programs but use more memory. will become
more valuable.

Just because structured programming concepts are usually expressed in high-level

languages does not mean that structured programming is not applicable to assembly
language programming. On the contrary. the assembly language programmer. with the
total freedom of expression that assembly level programming allows. needs the structUring concepts provided by structured programming. Creating modules with single
entry and exit POints. using simple control structures and keeping the compleXity of
each module minimal makes assembly language coding more efficient.

EXAMPLES
Response to a Switch
The structured version of this example is:

SWITCH = OFF
do while SWITCH = OFF
READ SWITCH
end
LIGHT = ON
DELAY 1
LIGHT = OFF

STRUCTURED
PROGRAMMING
IN THE
SWITCH AND
LIGHT SYSTEM

ON and OFF must have the proper definitions for the switch and light. We assume that
DELAY IS a module that provides a delay given by ItS parameter in seconds.
A statement In a structured program may actually be a subroutine. However. in order to
conform to the rules of structured programming. the subroutine cannot have any exits
other than the one that returns control to the main program.
Since "do-while" checks the condition before executing the loop. we set the variable
SWITCH to OFF before starting. The structured program IS straightforward. readable.
and easy to check by hand. However, It would probably require somewhat more memory than an unstructured program. which would not have to initialize SWITCH and could
combine the reading and checking procedures.

The Switch-Based Memory Loader

The switch-based memory loader is a more complex structured programming problem. We may implement the
flowchart of Figure 13-9 as follows (an indicates a commentl:

STRUCTURED
PROGRAMMING
FOR THE
SWITCH-BASED
MEMORY LOADER

INITIALIZE VARIABLES
HIADDRESS = 0
LOADDRESS = 0
THIS PROGRAM USES A DO-WHILE CONSTRUCT WITH NO CONDITION
(CALLED SIMPLY DO-FOREVER). THEREFORE. THE SYSTEM CONTINUALLY
EXECUTES THE PROGRAM CONTAINED IN THIS DO-WHILE LOOP
do forever

13-36

 TEST FOR HIADDRESS BUnON: PERFORM THE REQUIRED PROCESSING

IF IT IS ON.
if HIADDRBunON = 1 then
begin
HIADDRESS = SWITCHES
LIGHTS = SWITCHES
do
DELAY (DEBOUNCE TIME)
until HIADDRBUnON i= 1
end
TEST FOR LOADDRESS BUnON: PERFORM LOW ADDRESS PROCESSlf'lIG
IF IT is ON.
if LOADDRBUnON = 1 then
begin
LOADDRESS = SWITCHES
LIGHTS = SWITCHES
do
DELAY (DEBOUNCE TiME)
until LOADDRBUnON i= 1
end
TEST FOR DATABUnON. AND STORE DATA INTO MEMORY
IF IT IS ON.
if DATABUnON = 1 then
begin
DATA = SWITCHES
LIGHTS = SWITCHES
(HIADDRESS. LOADDRESSJ = DATA
do
DELAY (DEBOUNCE TIME)
until DATABUnON i= 1
end
end
THE LAST END ABOVE TERMINATES THE
do forever LOOP
Structured programs are not easy to write. but they can give a great deal of Insight into
the overall program logiC. You can check the logic of the structured program bv hand
before writing any actual code.

13-37

The Ctedit-Verification Terminal

Let us look at the keyboard entrY for the transaction terminal.
We will assume that the display array is ENTRY, the keyboard
strobe is KEYSTROBE. and the keyboard data is KEYIN, The struc
tured program without the function keys is:

NKEYS = 10

STRUCTURED
PROGRAM FOR
THE CREDIT
VERIFICATION
TERMINAL
STRUCTURED
KEYBOARD
ROUTINE

CLEAR ENTRY TO START

do while NKEYS > 0
NKEYS = NKEYS - 1
ENTRY(NKEYS) = 0
end
FETCH A COMPLETE ENTRY FROM KEYBOARD
do while NKEYS < 10
if KEYSTROBE = ACTIVE then
begin
KEYSTROBE = INACTIVE
ENTRY(NKEYS) = KEYIN
NKEYS = NKEYS + 1
end
end

Adding the SEND key means that the program must ignore extra digits after it has
a complete entry. and must ignore the SEND key until it has a complete entry. The
structured program is:

NKEYS = 10
CLEAR ENTRY TO START
do while NKEYS > 0
NKEYS = NKEYS 1
ENTRY(NKEYS) = 0
end
WAIT FOR COMPLETE ENTRY FOLLOWED BY SEND KEY
do while KEY i= SEND or NKEYS i= 10
if KEY STROBE = ACTIVE then
begin
KEYSTROBE = INACTIVE
KEY = KEYIN
if NKEYS i= 10 and KEY i= SEND then
begin
ENTRY(NKEYS) = KEY
NKEYS = NKEYS + 1
end
end
end

13-38

Note the following features of this structured program.

1)

The second if-then is nested within the first one. since keys are only entered after a
strobe is recognized. If the second if-then were on the same level as the first. a
single key could fill the entry. since its value would be entered into the arrav during
each iteration of the do-while loop.

2)

KEY need not be defined

ing of the entry.

In Itia IIv. ~ince

NKEYS IS set to zero as part of the clear-

Adding the CLEAR key allows the program to clear the entry originally by simulating the pressing of CLEAR: I.e.. by setting NKEYS to 10 and KEY to CLEAR before
starting. The structured program must also clear only digits that have previously been
filled. The new structured program is:
SIMULATE COMPLETE CLEARING
NKEYS = 10
KEY = CLEAR
WAIT FOR COMPLETE ENTRY AND SEND KEY
do while KEY #SEND or NKEYS # 10
CLEAR WHOLE ENTRY IF CLEAR KEY STRUCK
if KEY = CLEAR then
begin
KEY =0
do while NKEYS > 0
NKEYS = NKEYS - 1
ENTRY(NKEYS) = 0
end
end
GET DIGIT IF ENTRY INCOMPLETE
if KEYSTROBE = ACTIVE then
begin
KEYSTROBE = INACTIVE
KEY = KEYIN
if KEY < 10 and NKEYS # 10 then
begin
ENTRY(NKEYS) = KEY
NKEYS = NKEYS + 1
end
end
end
Note that the program resets KEY to zero after clearing the array. so that the operation IS
not repeated.

13-39

We can similarly build a structured program for the receive

routine. An initial program could look just for the header and
trailer characters. We will assume that RSTB is the indicator that a
character IS ready. The structured program is:
CLEAR HEADER FLAG TO START
HFLAG

=0

WAIT FOR HEADER AND TRAILER

do while HFLAG

= 0 or CHAR ~TRAILER

GET CHARACTER IF READY, LOOK FOR HEADER

if RSTB = ACTIVE then
begin
RSTB = INACTIVE
CHAR = INPUT
if CHAR = HEADER then HFLAG
end

=1

13-40

STRUCTURED
RECEIVE
ROUTINE

Now we can add the section that checks the message address against the three
digits in TERMINAL ADDRESS (TERMADDR). If any of the corresponding digits
are not equal, the ADDRESS MATCH flag (ADDRMATCH) is set to 1.
CLEAR HEADER FLAG. ADDRESS MATCH FLAG. ADDRESS COUNTER TO START
HFLAG = 0
ADDRMATCH = 0
ADDRCTR = 0
WAIT FOR HEADER. DESTINATION ADDRESS AND TRAILER
do while HFLAG = 0 or CHAR ;6TRAILER OR ADDRCTR;63
GET CHARACTER IF READY
if RSTB = ACTIVE then
begin
RSTB = INACTIVE
CHAR = INPUT
end
CHECK FOR TERMINAL ADDRESS AND HEADER
if HFLAG = 1 and ADDRCTR ;6 3 then
begin
ADDRMATCH
1
ADDRCTR = ADDRCTR + 1
end
if CHAR = HEADER then HFLAG = 1

end
The program must now walt for a header. a three-digit identification code. and a trailer.
You must be careful of what happens dUring the Iteration when the program finds the
header. and of what happens if an erroneous Identification code character is the same
as the trailer

13-41

A further addition can store the message in MESSG. NMESS is the number of
characters in the message: if it is not zero at the end. the program knows that the
terminal has received a valid message. We have not tried to minimize the logic expressions in this program.

CLEAR FLAGS. COUNTERS TO START

HFLAG =0
ADDRMATCH = 0
ADDRCTR =0
NMESS = 0
WAIT FOR HEADER. DESTINATION ADDRESS AND TRAILER
do while HFLAG = 0 or CHAR ~TRAILER or ADDRCTR ~3
GET CHARACTER IF READY
if RSTB = ACTIVE then
begin
RSTB = INACTIVE
CHAR = INPUT
end
READ MESSAGE IF DESTINATION ADDRESS = TERMINAL ADDRESS
if HFLAG = 1 and ADDRCTR = 3 then
if ADDRMATCH = 0 and CHAR ~TRAILER then
begin
MESSG(NMESS) = CHAR
NMESS = NMESS + 1
end
CHECK FOR TERMINAL ADDRESS
if HFLAG = 1 and ADDRCTR 1= 3 then
if CHAR ~TERMADDR(ADDRCTR) then
begin
ADDRMATCH = 1
ADDRCTR = ADDRCTR + 1
end
LOOK FOR HEADER
if CHAR = HEADER then HFLAG = 1
end

13-42

The program checks for the Identification code only if it found a header during a preVIOUS iteration. It accepts the message only if It has previously found a header and a
complete. matching destination address. The program must work properly during the
iterations when it finds the header. the trailer and the last digit of the destination address. It must not try to match the header with the terminal address or place the trailer
or the final digit of the destination address in the message. You might try adding the
rest of the logic from the flowchart (Figure 13-131 to the structured program. Note
that the order of operations is often critical. You must be sure that the program
does not complete one phase and start the next one during the same iteration.

REVIEW OF STRUCTURED PROGRAMMING

Structured programming brings discipline to program design. It forces you to limit
the types of structures you use and the sequence of operations. It provides singleentry, single-exit structures, which you can check for logical accuracy. Structured
programming often makes the designer aware of inconsistencies or possible combinations of inputs. Structured programming is not a cure-all, but it does bring
some order into a process that can be chaotic. The structured program should also
aid in debugging, testing, and documentation.
Structured programming is not simple. The programmer must not only define the
problem adequately, but must also work through the logic carefully. This is
tedious and difficult, but it results in a clearly written, working program.
TERMINATORS
The particular structures we have presented are not ideal and
are often awkward. In addition, it can be difficult to disFOR
tinguish where one structure ends and another begins, particSTRUCTURES
ularly if they are nested. Theorists may provide better structures in the future, or designers may wish to add some of their own. Some kind of
terminator for each structure seems necessary. since Indenting does not always clarify
the situation. "End" IS a logical terminator for the "do-while" loop. There IS no obVIOUS
terminator. however. for the "if-then-else" statement: some theOrists have suggested
"endif" or "fi" ("if" backwardS), but these are both awkward and detract from the
readability of the program.
We suggest the following rules for applying structured programming:
1)
2}

3)

RULES FOR
STRUCTURED
PROGRAMMING

Begin by writing a basic flowchart to help define the

logic of the program.
Start with the "sequential," "if-then-else:' and "do-while" constructs. They
are known to be a complete set. Le.. any program can be written In terms of these
structures.
Indent each level a few spaces from the previous level. so that you will know
which statements belong where.

4)

Use terminators for each structure; e.g.. "end" for the "do-while" and "endif" or
"fi" for the "if-then-else" The terminators plus the indentation should make the
program reasonably clear.

5}

Emphasize simplicity and readability. Leave lots of spaces. use meaningful

names. and make expressions as clear as possible. Do not try to minimize the logic
at the cost of clarity.
Comment the program In an organized manner.
Check the logic. Try all the extreme cases or special conditions and a few sample
cases. Any logical errors you find at this level will not plague you later.

6)
7)

13-43

TOP-DOWN DESIGN
The remaining problem is how to check and integrate modules
or structures. Certainly we want to divide a large task into
sub-tasks. But how do we check the sub-tasks in isolation and
put them together? The standard procedure, called "bottom-up design," requires
extra work in testing and debugging and leaves the entire integration task to the
end. What we need is a method that allows testing and debugging in the actual
program environment and modularizes system integration.
This method is "top-down design." Here we start by writing
the overall supervisor program. We replace the undefined subprograms by program "stubs," temporary programs that may
either record the entry, provide the answer to a selected test
problem, or do nothing. We then test the supervisor program
to see that its logic is correct.

TOP-DOWN
DESIGN
METHODS
STUBS

We proceed by expanding the stubs. Each stub will often conEXPANDING

tain sub-tasks, which we will temporarily represent as stubs.
STUBS
This process of expansion, debugging, and testing continues
ADVANTAGES
until all the stubs are replaced by working programs. Note that
OF TOP-DOWN
testing and integration occur at each level. rather than all at the
DESIGN
end. No special dnver or data generation programs are necessary.
We get a clear idea of exactly where we are in the design. Topdown design assumes modular programming, and is compatible with structured
programming as well.
The disadvantages of top-down design are:
1)

The overall design may not mesh well with system hardware.

DISADVANTAGES
OF TOP-DOWN
DESIGN

2)

It may not take good advantage of existing software.

3)

Stubs may be difficult to write, particularly if they must

work correctly In several different places.

4)
5)

Top-down design may not result In generally useful modules.

Errors at the top level can have catastrophic effects, whereas errors In bottom-up
design are usually limited to a particular modu Ie.

In large programming projects, top-down design has been shown to greatly improve programmer productivity. However, almost all of these projects have used
some bottom-up design in cases where the top-down method would have
resulted in a large amount of extra work.
Top-down design is a useful tool that should not be followed to extremes. It provides the same discipline for system testing and integration that structured programming provides for module design. The method, however, has more general
applicability, since it does not assume the use of programmed logic. However,
top-down design may not result in the most efficient implementation.

13-44

EXAMPLES
Response to a Switch
The first structured programming example actually demonstrates top-down design as well. The program was:
SWITCH = OFF
do while SWITCH = OFF
READ SWITCH
end
LIGHT = ON
DELAY 1
LIGHT = OFF

TOP-DOWN
DESIGN
OF SWITCH
AND LIGHT
SYSTEM

These statements are really stubs. since none of them is fully defined. For example. what does READ SWITCH mean? If the switch were one bit of input port SPORT. it
really means:
SWITCH

= SPORT AND SMASK

where SMASK has a '1' bit in the appropriate position. The masking may. of course. be
implemented with a Bit Test Instruction.
Similarly. DELAY 1 actually means (if the processor Itself provides the delay):
REG = COUNT
do while REG ~O
REG = REG - 1
end
COUNT is the appropriate number to provide a one-second delay. The expanded version of the program is:
SWITCH =0
do while SWITCH = 0
SWITCH = SPORT AND MASK
end
LIGHT = ON
REG = COUNT
do while REG ~O
REG = REG - 1
end
LIGHT = NOT (LIGHT)

Certainly this program is more explicit, and could more easily be translated into
actual instructions or statements.

13-45

The Switch-Based Memory Loader

This example is more complex than the first example. so we
must proceed systematically. Here again. the structured program contains stubs.
For example. if the HIGH ADDRESS button IS one bit of input
port CPORT. "if HIADDRBUTTON = 1" really means:
1)

Input from CPORT

2)
3)

Complement
Logical AND with HAM ASK

TOP-DOWN
DESIGN OF
SWITCH-BASED

MEMORY
LOADER

where HAMASK has a '1' in the appropriate bit POSItIOn and 'Os' elsewhere. Similarly
the condition "if DATABUTTON = 1" really means:
1)

Input from CPORT

2)

Complement

3)

Logical AND with DAMASK

So. the,lnitial stubs could lust assign values to the buttons. e.g"
HIADDRBUTTON = 0
LOADDRBUTTON = 0
OATABUTTON = 0
A run of the supervisor program should show that it takes the implied "else" path
through the "if-then-else" structures. and never reads the SWitches. Similarly. if the
stub were:
HIADDRBUTTON

=1

the supervisor program should stay in the "do while HIADDRBUTTON = 1" loop waitIng for the button to be released. These simple runs check the overall logiC.
Now we can expand each stub and see if the expansion produces a reasonable
overall result. Note "ow debugging and testing proceed in a straightforward and
modular manner. We expand the HIADDRBUTTON = 1 stub to:
READ CPORT
HIADDRBUTTON = NOT (CPORT) AND HAMASK
The program should wait for the HIGH ADDRESS button to be closed. The program
should then display the values of the SWitches on the lights. This run checks for the
proper response to the HIGH ADDRESS button.
We then expand the LOW ADDRESS button module to:
READ CPORT
LOADDIi13UTTON = NOT (CPORT) AND LAMASK
With the LOW ADDRESS button in the closed position. the program should display the
values of the switches on the lights. This run checks for the proper response to the LOW
ADDRESS button.
Similarly. we can expand the DATA button module and check for the proper response
to that button. The entire program will then have been tested.

When all the stubs have been expanded. the coding, debugging, and testing
stages will all be complete. Of course, we must know exactly what results eac"
stub should produce. However, many logical errors will become obvious at each
level without any further expansion.

13-46

Kevboard

ACK =0

No

TransmIt

Displav

Receive

Figure 13-19 Initial Flowchart for Transaction Terminal

The Transaction Terminal

This example. of course. will have more levels of detail. We
cou Id start with the following program (see Figu re 13-19 for
a flowchart):
KEYBOARD
ACK =0
do while ACK = 0
TRANSMIT
RECEIVE
end
DISPLAY

TOP-DOWN
DESIGN OF
VERIFICATION
TERMINAL

Here KEYBOARD, TRANSMIT, RECEIVE, and DISPLAY are program stubs that will
be expanded later. KEYBOARD. for example. could slmplv place a ten-digit verified
number into the appropriate buffer.

13-47

VER =0

No

Complete =0

Yes

KEVIN
KEVOS

Verify

Figure 13-20. Flowchart for Expanded KEYBOARD Routine

The next stage of expansion could produce the following program for KEYBOARD (see Figure 13-201:
VER = 0
do while VER = 0
COMPLETE =0
do while COMPLETE = 0
KEYIN
KEYDS
end
VERIFY
end

EXPANDING
THE
KEYBOARD
ROUTINE

Here VER = 0 means that an entry has not been verified: COMPLETE = 0 means that
the entry is incomplete. KEYIN and KEYDS are the keyboard Input and display routines
respectively VERIFY checks the entry. A stub for KEYIN would simply place a random
entry (from a random number table or generator) Into the buffer and set COMPLETE to
1.

We would continue by similarly expanding. debugging. and testing TRANSMIT.

RECEIVE. and DISPLAY. Note that you should expand each program by one level
so that you do not perform the integration of an entire program at anyone time.
You must use your judgment in defining levels. Too small a step wastes time,
while too large a step gets you back to the problems of system integration that
top-down design is supposed to solve.
13-48

REVIEW OF TOP-DOWN DESIGN

Top-down design brings discipline to the testing and integration stages of program design. It provides a systematic method for expanding a flowchart or problem definition to the level required to actually write a program. Together with
structured programming, it forms a complete set of design techniques.
Like structured programming, top-down design is not simple. The designer must
have defined the problem carefully and must work systematically through each
level. Here again the methodology may seem tedious, but the payoff can be substantial if you follow the rules.
We recommend the following approach to top-down design:

FORMAT
FOR
TOP-DOWN
DESIGN

1)
2)

Start with a basic flowchart.

Make the stubs as complete and as separate as possible.

3)
4)

Define precisely all the possible outcomes from each stub

and select a test set.
Check each level carefully and systematically.

5)
6)

Use the structures from structured programming.

Expand each stub by one level. Do not try to do too much

7)
8l

Watch carefully for common tasks and data structures.

Test and debug after each stub expansion. Do not try to do an entire level at a
time.
Be aware of what the hardware can do. Do not hesitate to stop and do a little
bottom-up design where that seems necessary.

9)

In

one step.

REVIEW OF PROBLEM DEFINITION AND PROGRAM DESIGN

You should note that we have spent an entire chapter without mentioning any
specific microprocessor or assembly language, and without writing a single line of
actual code. Hopefully, though, you now know a lot more about the examples than
you would have if we had just asked you to write the programs at the start.
Although we often think of the writing of computer instructions as a key part of
software development, it is actually one of the easiest stages.
Once you have written a few programs, coding will become simple. You will soon
learn the instruction set, recognize which instructions are really useful, and
remember the common sequences that make up the largest part of most programs. You will then find that many of the other stages of software development
remain difficult and have few clear rules.
We have suggested here some ways to systematize the important early stages. In
the problem definition stage, you must define all the characteristics of the
system - its inputs, outputs, processing, time and memory constraints, and error
handling. You must particularly consider how the system will interact with the
larger system of which it is a part, and whether that larger system includes
electrical equipment, mechanical equipment, or a human operator. You must start
at this stage to make the system easy to use and maintain.
In the program design stage, several techniques can help you to systematically
specify and document the logic of your program. Modular programming forces you
to divide the total program into small, distinct modules. Structured programming
provides a systematic way of defining the logic of those modules, while top-down
design is a systematic method for integrating and testing them. Of course, no one
can compel you to follow all of these techniques; they are. in fact, guidelines more
than anything else. But they do provide a unified approach to design, and you
should consider them a basis on which to develop your own approach.

13-49

REFERENCES
1.

See. for example. V. P Srinl. "Fault DiagnosIs of Microprocessor Systems." Computer. January 1977. pp. 60-65. For a description of signature analysis. see G. Gordon and H. Nadig. "Hexadecimal Signatures Identify Trouble-spots In
Microprocessor Systems." Electronics. March 3. 1977. pp. 89-96. There IS also an
Application Note (#222) entitled "A DeSigner's GUide to Signature AnalysIs"
available from Hewlett-Packard.

2.

For a brief diSCUSSion of human factors conSiderations. see G. Morns. "Make Your
Next Instrument DeSign Emphasize User Needs and Wants." EON. October 20.
1978. pp. 100-105.
-

3.

D. L. Parnas (see the references below) has been a leader In the area of modular programming.

4.

Collected by B. W. Unger (see reference below!.

5.

Formulated by D. L. Parnas.

The follOWing references prOVide additional information on problem definition and program design:
Chaplin. N.. Flowcharts. Auerbach. Princeton. N. J.. 1971.
Dahl. O. J.. C. A. R. Hoave. and E. W. Dijkstra. Structured Programming. Academic
Press. New York. N. Y.. 1972.
Dalton. W. F.. "DeSign Microcomputer Software like Other Systems Electronics. January 19. 1978. pp. 97-101.
Dijkstra. E.
1976.

w..

Systematically."

A Discipline of Programming. Prentice-Hall. Englewood Cliffs. N. J..

Halstead. M. H.. Elements of Software SCience. American ElseVier. New Yark. 1977
Hughes. J. K. and J. I. Michtom. A Structured Approach to Programming. Prentice-Hall.
Englewood Cliffs. N. J.. 1977
Morgan. D. E. and D. J. Taylor. "A Survey of Methods for Achieving Reliable Software."
Computer. February 1977. pp. 44-52.
Myers. W.. "The Need for Software Engineering." Computer. February 1978. pp. 12-25.
Parnas. D. L.. "On the Criteria to be Used In Decomposing Systems Into Modules." Communications ~ the ACM. December 1972. pp. 1053-1058.
-Parnas. D. L.. "A Technique for the Specification of Software Modules with Examples."
Communications ! the ACM. May 1973. pp. 330-336.
Schneider. V.. "Prediction of Software Effort and Prolect Duralion - Four New Formulas." SIGPLAN Notices. June 1978. pp. 49-59.
Shnelderman. B. et al.. "Experimental Investigations of the Utility of Detailed Flowcharts in Programming." CommunicatiOns ~ the ACM. June 1977. pp. 373-381.
Ulrickson. R. W .. "Software Modules Are the Building Blocks." Electronic Design.
February 1. 1977. pp. 62-66.
Ulrickson. R. W .. "Solve Software Problems Step-by-Step." Electronic Design. January
18.1977. pp. 54-58.
-Unger. B. W" "Programming Languages for Computer System SimulatiOn." Simulation.
April 1978. pp. 101-110.

13-50

Wirth. N.. Algorithms + Data Structures

N. J.. 1976.

= Programs.

Prentice-Hail. Englewood Cliffs.

Wirth. N.. Systematic Programming: an Introduction. Prentice-Hail. Englewood Cliffs. N.

J.. 1973.
Yourdon. E. U.. Techniques of Program Structure and Design. Prentice-Hall. Englewood
Cliffs. N. J .. 1975.

13-51

Chapter 14
DEBUGGING AND TESTING
As we noted at the beginning of the previous chapter. debugging and testing are
among the most time-consuming stages of software development. Even though such
methods as modular programming. structured programming. and top-down design
can simplify programs and reduce the frequency of errors. debugging and testing
still are difficult because they are so poorly defined. The selection of an adequate set
of test data is seldom a clear or sCientific process. Finding errors sometimes seems like a
game of "Pin the tail on the donkey." except that the donkey IS moving and the programmer must position the tail by remote control. Surely. few tasks are as frustrating as
debugging programs.

This chapter will first describe the tools available to aid in debugging. It will then
discuss basic debugging procedures. describe the common types of errors. and
present some examples of program debugging. The last sections will describe
how to select test data and test programs.
We will not do much more than describe the purposes of most of the debugging tools.
There is very little standardization In this area. and not enough space to discuss all the
devices and programs that are currentlv available. The examples should give you some
Idea of the uses, advantages, and limitations of particular hardware or software aids.

SIMPLE DEBUGGING TOOLS

The simplest debugging tools available are:
A single-step facility
A breakpoint facility
A Register Dump program (or utility)
A Memory Dump program
The single-step facility allows you to execute the program one
step at a time. Most ZSO-based microcomputers have thiS facility.
since the circuitry IS fairly simple. Of course. the only things that
you will be able to see when the computer executes a single-step are the states
of the output lines that you are monitoring. The most important lines are:
Data Sus
Address Bus
Control lines MREO (Memory Requestl. 10RO (Input/Output Request), RD (Memory
Read), and WR (Memory Write).

If you monitor these lines (either in hardware or in software). you will be able to
see the progression of addresses. instructions. and data as the program executes.
You will be able to tell what kind of operations the CPU is performing. ThiS information will Inform you of such errors as Incorrect Jump instructions. omitted or incorrect addresses. erroneous operation codes. or Incorrect data values. However. you cannot see the contents of registers and flags Without some additional debugging facility
or a speCial sequence of instructions. Many of the operations of the program cannot be
checked In real time.

14-1

Table 14-1

Z80 Restart and Interrupt Addresses

Instruction or External Input

(Mnemonic!
(Pinl
RST
RST
RST
RST
RST
RST
RST
RST

OOH
08H
10H
18H
20H
28H
30H
38H or INT In Mode 1
NMI

Instruction Oblect Code

(Hex)

Destination Address
(Hex)

C7

0000
0008
0010
0018
0020

CF

07
OF

E7

EF

0028

F7
FF

0030
0038
0066

LIMITATIONS
There are many errors that a single-step mode cannot help you
OF SINGLEto find. These Include timing errors and errors In the Interrupt
STEP MODE
or DMA systems. Furthermore. the single-step mode IS very
slow. typically executing a program at less than one millionth
of the speed of the processor Itself. To single-step through one second of real processor
time would take more than ten days. The single-step mode IS useful only to check the
logic of short Instruction sequences.

A breakpoint is a place at which the program will automatBREAKPOINT)

ically halt or wait so that the user can examine the current
status of the system. The program will usually not start again until the operator requests a resumption of execution. Breakpoints allow you to check or pass through an
entire section of a program. Thus. to see if an initialization routine IS correct. you can
place a breakpOint at the end of It and run the program. You can then check memory
locations and registers to see if the entire section IS correct. However. note that if the
section IS not correct. you'lI still have to Pin down the error. either With earlier breakPOints or With a Single-step mode.
Breakpoints complement the single-step mode. You can use breakpoints either to
localize the error or to pass through sections that you know are correct. You can
then do the detailed debugging in the single-step mode. In some cases. breakpoints
do not affect program timing; they can then be used to check Input/output Interrupts.
Breakpoints often use part or all of the microprocessor Interrupt
system. Some microprocessors have a speCial SOFTWARE INTERRUPT or TRAP facility that can act as a breakpoint. On the Z80, if
you are not already using all the RST vectors in your program, you can use the RST
(Restart) instruction as a breakpoint. Table 14-1 gives the destination addresses for
the vanous RST Instructions. Chapter 12 describes the RST Instruction In more detail.
The breakpoint routine can pnnt register and memory contents or iust wait (e.g.. execute HALT or a conditional Jump dependent on a switch input) until you allow the
computer to proceed. If you are not uSing the maskable Interrupt (lNT) or the nonmaskable Interrupt (NMI) In your system. you can use those vectors as externally controlled breakpoints. But remember that the interrupts (including NMI) and RST use the
Stack and Stack Pointer to store the return address. Figure 14-1 shows a routine where
RST results in an endless loop. You would have to clear this breakpoint with a RESET or
Interrupt signal.

14-2

Figure 14-1. A Simple Breakpoint Routine

RST1B

lBH
lBH
RST1B

ORG
EQU
JR

WAIT IN PLACE

The simplest method for inserting breakpoints is to replace the first byte of the instruction with a RST instruction or to replace the instruction with a Jump or CALL
instruction. Use of a RST instruction IS preferred on the
since It Involves the
replacement of only a single byte. whereas a JP or CALL Involves three bytes. The JR
instruction IS not sUitable for breakpointlng because you cannot guarantee that the
debug software IS within -126 to + 129 bytes of the instruction being breakpointed.
Multiple-byte Instructions used to Implement breakpoints can cause problems on the
due to the presence of single-byte instructions. To illustrate this program. examine
the program segment shown below:

zao.

zao

Memory Address
(Hex)

Memory Contents
(Hex)

100
101
102

7B
87
87

Instruction
(Mnemonic)

L1
L2:

LD
ADD
ADD

A.E
A.A
AA

If you wish to set a breakpoint at location 10016 using a 3-byte CALL or JP. the code at
locations 10116 and 10216 will also get overlaid by the CALL or JP instruction. This
means that the debugger has to be aware that these locations have also been modified.
Any transfers of control to L1 or L2 while the breakpoint IS set will produce unexpected
results unless the debugger IS designed to catch this case. This added complexity can
be avoided by uSing a RST instruction.
Many monitors have facilities for inserting and removing
INSERTING
BREAKPOINTS
breakpoints implemented via some type of Jump Instruction.
Such breakpoints do not affect the timing of the program until
the breakpoint IS executed. However. note that this procedure will not work if part or all
of the program IS In ROM or PROM. Other monitors Implement breakpoints by actually
checking the address lines or the Program Counter In hardware or In software. ThiS
method allows breakpOints on addresses In ROM or PROM. but It may affect the timing
if the address must be checked In software. A more powerful facility would allow the
user to enter an address to which the processor wou Id transfer control. Another
possibility would be a return dependent on a switch:
RST18
WAITS;

ORG
EQU
PUSH
IN
BIT
JR
POP
RET

18H
18H
AF
A.(PIODRA)
SW.A
NZWAITS
AF

:SAVE ACCUMULATOR. FLAGS

:GET SWITCH DATA
:IS SWITCH CLOSED?
:NO. WAIT UNTIL IT IS
:RESTORE ACCUMULATOR. FLAGS

Remember to re-enable the interrupts if the routine uses an external interrupt input.

14-3

Store all registers

in Stack
COUNT

= Number of

bytes In register = 22
Data POinter =
Stack Pointer + 20

Store Data POinter

In Stack

Data Pointer ==
Data POinter - 1

Print (Data POinter)

as 2 hex digits
= COUNT -1

COUNT

Restore all registers

from Stack

Figure 14-2. Flowchart of Register Dump Program

A Register Dump utility on a microcomputer is a program that

lists the contents of all the CPU registers. ThiS Information is
usually not directly obtainable. The following routine will print
the contents of all the registers on the system printer. if we assume that PRTHEX
prints the contents of the Accumulator as two hexadecimal digits. Figure 14-2 IS a
flowchart of the program and Figure 14-3 shows a typical result. We assume that the
routine IS entered With a CALL instruction that stores the old Program Counter at the
top of the Stack.

14-4

 PLACE ALL CPU REGISTER CONTENTS IN STACK (PC ALREADY ON STACK)

PUSH
PUSH
PUSH
PUSH
PUSH
PUSH
EX
EXX
PUSH
PUSH
PUSH
PUSH

AF
BC
DE
HL
IX
IY
AF.AF

;SAVE REGULAR USER REGISTERS

;SAVE INDEX REGISTERS

;ACCESS AND SAVE PRIMED CPU REGISTERS

AF
BC
DE
HL

USE STACK POINTER AS STARTING ADDRESS

LD
ADD
LD
ADD
PUSH

HL.O
HL.SP
DE.20
HL.DE
HL

;GET STACK POINTER

;COMPUTE ORIGINAL STACK POINTER
;SAVE ORIGINAL STACK POINTER IN STACK

PRINT CONTENTS OF REGISTERS

ORDER IS PC(HIGHl.PC(LOWlAF.B.C.D.E.H.L.IX(HIGHl.IX(LOWl.IY(HIGHl.
IY (LOWl,A.F .S. C'. O' .E'.H .L'.SP(HIGHl,SP(LOW)

PRNTl

LD
DEC
LD
CALL
DJNZ

B.22
HL
A,(HU
PRTHEX
PRNTl

;NUMBER OF BYTES = 22
:GET A BYTE FROM STACK
:AND PRINT IT

, RESTORE REGISTERS FROM STACK

POP
POP
POP
POP
POP
EX
EXX
POP
POP
POP
POP
POP
POP
RET

HL
HL
DE
BC
AF
AF.AF

:POP AND DISCARD ORIGINAL STACK POINTER

;RESTORE PRIMED CPU REGISTERS

IY
IX
HL
DE
BC
AF

:RESTORE INDEX REGISTERS

;RESTORE REGULAR CPU REGISTERS

14-5

10
42
07

(A)
(F)
(B)

3E
23

(C)
(D)

01
17
01
D3

(E)
(H)
(L)

(IX)

58
E2

(lY)

A2
36

(A')

67

(F)

E8
11

(B')
(C')

EB

(D')

09
D7
66

(E')
(H')
(L')
(STACK POINTER)

68
E2

Figure 14-3 Results of a Typical Z80 Register Dump

14-6

A Memory Dump is a program that lists the contents of memory on an output device (such as a printer). This is a much more
efficient way to examine data arrays or entire programs than Just
looking at single locations. However. very large memory dumps are not useful (except
to supply scrap paper) because of the sheer mass of information that they produce.
They may also take a long time to execute on a slow printer. Small dumps may.
however, provide the programmer with a reasonable amount of information that
can be examined as a unit. Relationships such as regular repetitions of data patterns or offsets of entire arrays may become obvious.
A general dump IS often rather difficult to write. The programmer should be careful of
the following situations:
1)
2)

The size of the memory area exceeds 256 bytes. so that an 8-bit counter will not
suffice.
The ending location IS an address smaller than the starting location. This can be
treated as an error. or simply cause no output. since the user would seldom want to
print the entire memory contents In an unusual order.

Since the speed of the Memory Dump depends on the speed of the output device, the
effiCiency of the routine seldom matters. The following program will ignore cases
where the starting address is larger than the ending address. and will handle
blocks of any length. We assume that the starting address is in Register Pair DE and
the ending address IS in Register Pair HL.
. STOP IF ENDING ADDRESS BEFORE STARTING ADDRESS
AND
SBC

JR

HL,DE
C.DONE

XCHG
INC

DE

:CLEAR CARRY
:IS ENDING ADDRESS BEFORE STARTING?
:YES. DO NOT DUMP ANYTHING
:GET STARTING ADDRESS INTO HL
:COUNT = NUMBER OF LOCATIONS TO BE
DUMPED

PRINT CONTENTS OF LOCATIONS

DUMP:

DONE:

LD
CALL
INC
DEC
LD
OR

A.(HLl
PRTHEX
HL
DE

:GET CONTENTS OF A LOCATION

:AND PRINT IT

A.E

:ALL LOCATIONS DUMPED?

JR

NZ,DUMP

:NO. CONTINUE DUMPING

HALT

Note that the onlY 16-blt Subtract instruction IS SSC. which subtracts the contents of a
register pair and the Carry from Register Pair HL. SBC. like other Subtract Instruclions,
sets the Carry if a borrow is required (contrary to what some Z80 manuals say).

Figure 14-4 shows the output from a dump of memory locations 1000 to 101F.

14-7

23
6E
47
34

IF
42
36
ED

60
38
23
BC

54
17
81
AF

37
59
El
FE

28
44
FF
FF

3E
98
FF
27

00
37
5A
02

Figure 14-4. Results of a Typical Memory Dump

This routine correctly handles the case In which the starting and ending locations are
the same (try It I). You will have to Interpret the results carefully if the dump area
includes the Stack. since the dump subroutine itself uses the Stack. PRTHEX may also
change memory and Stack locations.
In a memory dump. the data can be displayed In a number of different ways. Common
forms are ASCII characters or pairs of hexadeCimal digits for 8-bit values and four hexadecimal digits for 16-blt values. The format should be chosen based on the Intended
use of the dump. It is almost always easier to Interpret an oblect code dump if It IS displayed In hexadeCimal form rather than ASCII form.
A common and useful dump format is illustrated here:
1000546865206475 6D 70

The dump

Each line consists of three parts. The line starts With the hexadeCimal address of the
first byte displayed on the line. Following the address are eight or sixteen bytes displayed In hexadecimal form. Last IS the ASCII representation of the same eight or SIXteen bytes. Try rewriting the memory dump program so that it will print the address and
the ASCII characters as well as the hexadeCimal form of the memory contents.

MORE ADVANCED DEBUGGING TOOLS

The more advanced debugging tools that are most widely used are:
Similar programs to check software
Logic analyzers to check signals and timing
Many Variations of both these tools eXiSt. and we shall discuss only the standard
features.

The simulator is the computerized eqUivalent of the pencil-andpaper computer. It is a computer program that goes through the
operating cycle of another computer. keeping track of the contents of all the registers, flags, and memory locations. We could. of course. do thiS
by hand. but It would require a large amount of effort and close attentIOn to the exact
effects of each instruction. The slmu lator program never gets tired or confused. forgets
an instruction or register. or runs out of paper.
Most Simulators are large FORTRAN programs. They can be purchased or used on the
time-sharing serVices. The Z80 Simulator IS available In several versions from different
sources.

14-8

Typical simulator features are:

1)

A breakpoint facility. Usually. breakpoints can be set after a particular number of

cycles have been executed. when a memory location or one of a set of memory
locations is referenced. when the contents of a location or one of a set of locations
are altered. or on other conditions.

2)

Register and memory dump facilities that can display the values of memory locations. registers. and I/O ports.

3)

A trace facility that will print the contents of particular registers or memory locations whenever the program changes or uses them.

4)

A load facility that allows you to set values initially or change them dUring the
simulation.

Some simulators can also simulate Input/output. Interrupts. and even DMA.

The simulator has many advantages:

1)

It can provide a complete deSCription of the status of the computer. since the
Simulator program is not restricted by Pin limitations or other characteristics of the
underlYing circuitry.

2)

It can proVide breakpoints. dumps. traces. and other facilities. Without using any of
the processor's memory space or control system. These facilities will therefore not
Interfere with the user program.

3)

Programs. starting POints. and other conditions are easy to change.

4)

All the facilities of a large computer. including peripherals and software. are available to the microprocessor deSigner.

On the other hand, the simulator is limited by its software base and its separation
from the real microcomputer. The major limitations are:
1)

The Simulator cannot help With timing problems. since it operates far more slowly
than real time and does not model actual hardware or Interfaces.

2)

The simulator cannot fully model the input/output section.

3)

The Simulator is usually qUite slow. Reproducing one second of actual processor
time may require hours of computer time. Using the simulator can be qUite expensive.

The simulator represents the software side of debugging; it has the typical advantages and limitations of a wholly software-based approach. The simulator can
provide insight into program logic and other software problems, but cannot help
with timing, I/O, and other hardware problems.
The logic or microprocessor analyzer is the hardware solution
to debugging. Basically, the analyzer is the parallel digital version of the standard oscilloscope. The analyzer displays information In binary. hexadecimal or mnemonic form on a CRT. and has a variety of triggering
events. thresholds. and Inputs. Most analyzers also have a memory so that they can display the past contents of the busses.
The standard procedure IS to set a triggering event. such as the occurrence of a particular address on the Address Bus or instruction on the Data Bus. For example. one might
trigger the analyzer if the microcomputer tries to store data in a particular address or execute an Input or output Instruction. One may then look at the sequence of events that
preceded the breakpOint. Common problems you can find in this way include short
noise spikes (or glitches). incorrect signal sequences. overlapping wave-forms.
and other timing or signaling errors. Of course, a software simulator could not be
used to diagnose those errors any more than a logic analyzer could conveniently
be used to find errors in program logic.

14-9

Logic analyzers vary in many respects. Some of these are:

IMPORTANT

21

FEATURES
Number of Input lines. At least 24 are necessary to monitor
OF LOGIC
an 8-blt Data Bus and a 16-blt Address Bus. Still more are
ANALYZERS
necessary for control signals. clocks. and other Important inputs.
Amount of memory. Each previous state that IS saved will occupy several bytes.

3)

Maximum frequency. It must be several MHz to handle the fastest processors.

1)

4)

Minimum signal width (important for catching glitches),

51

Type and number of triggering events allowed. Important features are pre- and
post-trigger delays; these allow the user to display events occurring before or
after the trigger event.
Methods of connecting to the microcomputer. This may require a rather complex
Interface.
Number of display channels.

6)
7}
8)

Binary. hexadecimal or mnemonic displays.

9)

Display formats.

10)

Signal hold time requirements.

111
12)

Probe capacitance.
Single or dual thresholds

All of these factors are Important In comparing different logic and microprocessor
analyzers. since these instruments are new and unstandardized. A tremendous variety
of products IS already available and this variety will become even greater In the future.

Logic analyzers, of course. are necessary only for systems with complex timing.
Simple applications with low-speed peripherals have few hardware problems that
a designer cannot handle with a standard oscilloscope.

DEBUGGING WITH CHECKLISTS

The designer cannot possibly check an entire program by hand; however. there are
certain trouble spots that the designer can easily check. You can use systematic hand
checking to find a large number of errors without resorting to any debugging tools.

The question is where to place the effort. The answer is on

WHAT TO
INCLUDE IN
points that can be handled with either a yes-no answer or with
a simple arithmetic calculation. Do not try to do complex
CHECKLIST
arithmetic. follow all the flags. or try every conceivable case. limit
your hand checking to matters that can be settled easily. Leave the complex problems
to be solved With the aid of debugging tools. But proceed systematically: build your
checklist. and make sure that the program performs the basIc operations correctly.
The first step is to compare the flowchart or other program documentation with
the actual code. Make sure that everything that appears in one also appears in the
other. A simple checklist will do the lob. It is easy to completely omit a branch or a processing section.
Next concentrate on the program loops. Make sure that all registers and memory
locations used inSide the loops are initialized correctly. This is a common source of errors; once again. a simple checklist will suffice.
Now look at each conditional branch. Select a sample case that should produce a
branch and one that should not: try both of them. Is the branch correct or reversed? If
the branch Involves checking whether a number IS above or below a threShold. try the
equality case. Does the correct branch occur? Make sure that your chOice IS consistent
with the problem definllion.

14-10

Look at the loops as a whole. Try the tirst and last Iterations by hand: these are often
troublesome special cases. What happens if the number ot Iterations IS zero; I.e., there
IS no data or the table has no elements? Does the program tall through correctly? Programs often will perform one Iteration unnecessarily. or. even worse. decrement counters past zero before checking them.
Check off everything down to the last statement. Don't assume (hopefully) that
the first error is the only cine in the program. Hand checking will allow you to get
the maximum benefit from debugging runs, since you will get rid of many simple
errors ahead of time.
A quick review of the hand checking questions:
1)

Is every element of the program deSign In the program (and

vice versa for documentation purposes)?

2)

Are all registers and memory locations used inside loops inItialized before they are used?

3)

Are all conditional branches correct?

Do all loops start and end properly?

4)
5)

Are equality cases handled correctly?

6)

Are triVial cases handled correctly?

HAND
CHECKING
QUESTIONS

LOOKING FOR ERRORS

Of course, despite all these precautions (or if you skip over
some of them), programs often still don't work. The designer
is left with the problem of how to find the mistakes. The hand
checklist provides a starting place if you didn't use it earlier; some of the errors
that you may not have eliminated are:
1)

Failure to initialize variables such as counters, pointers, sums, etc. Do not

assume that registers. memory locations. or flags necessarily contain zero before
they are used.

2)

Inverting the logic of a conditional jump, such as uSing Jump on Carry when you
mean Jump on Not Carry. Remember the effects of a comparison or subtraction (A
IS the contents of the Accumulator. M the contents of the register or memory location):
Zero flag

1 if A =M

Carry flag

o if A ~M
1 if A < M

o if A ~

Note particularly that Carry = 0 if A = M. (the equality case). So. Jump on Carry
means Jump if A < M. and Jump on Not Carry means lump if A 2:: M. If you want
the equality case on the other Side. try either reversing the roles of A and M or
adding 1 to M. For example. if you want a IU mp if A ~ 10. use:
CP
JR

10
NC.ADDR

If. on the other hand. you want a lump It A

CP
JR
3)

> 10.

use:

11
NC.ADDR

Updating the counters and pointers in the wrong place or not at all. Be sure
that there are no paths through a loop that either skip or repeat the updating instructions.

14-11

4)

Failure to fall through correctly in trivial cases such as no data In a buffer. no

tests to be run, or no entnes In a transaction. Do not assume that such cases will
never occur unless the program specifically eliminates them.

Other problems to watch for are:

5)

Reversing the order of operands. Remember that the LD instruction moves the
second operand into the first operand, For example, LD B,A moves the contents of
A to B, not the other way around.

6)

Changing condition flags before you use them.

Remember that INC and DEC, when applied to a single register or memory location, affect all the flags except Carry. Remember also that POP AF and EX AF.AF'
affect all the flags, and that Logical instructions clear the Carry.

7)

Failing to change condition flags when you intend to.

The Zero and Sign flags may not represent the current state of the Accumulator.
since many Instructions (particularly LD) do not change the flags. Note that Incremenling or decrementing register pairs (for example, INC HL or DEC BC) and complementing the Accumulator (CPU affect no flags at all.

8)

Confusing values and addresses.

Remember that LD HL.l000H loads HL with the number 1000 (hex) while LD
HL.(1000H) loads HL with the contents of memory locations 1000 and 1001. A
similar distinction applies to LD ACOUNT and LD A.(COUNT).

9)

Accidentally reinitializing a register or memory location.

Make sure that no Jump Instructions transfer control back to initialization statements.

10)

Confusing numbers and characters.

Remember that the ASCII and EBCDIC representations of digits differ from the
digits themselves. For example, ASCII 7 IS hex 37. whereas hex 07 IS the ASCII
BELL character.

111

Confusing binary and decimal numbers.

Remember that the BCD representation of a number differs from its binary representation. For example, BCD 36. when treated as a simple hexadecimal constant.
is equlvalent to 54 deCimal (try itl.

12)

Reversing the order in subtraction. Be careful also with other operations (like
division) that do not commute. Remember that SUB and CP produce A-M. not
M-A.

13)

Ignoring the effects of subroutines and macros.

Don't assume that calls to subroutines or invocations of macros will not change
flags. registers, or memory locations. Be sure of exactly what effects subroutines
or macros have. Note that It IS very Important to document these effects so that
the user can determine them without gOing through the entire listing.

141

Using the Shift instructions improperly.

Remember the precise effects of RLC, RL, RRC. RR. SLA. SRA. and SRL. They are
all 1-blt shifts. SLA and SRL both clear the empty bit. SRA preserves the sign
(most significant bill by extending It to the nght. RLC and RRC are circular shifts
that do not Include the Carry In the circular register; RL and RR are circular shifts
that Include the Carry. Remember that these instructions affect all the flags, even
if they are applied to the data in a memory location. Note, however. that the oneword shifts RLCA, RLA. RRCA, and RRA affect only the Carry.

14-12

15)

Counting the length of an array incorrectly.

Remember that there are five (not fourl memory locations Included In addresses
0100 through 0104. inclusive.

16)

Confusing registers and register pairs.

Remember that the CPU registers and register pairs are physically the same. You
can use them singly for 8-bll data or in pairs for addresses or 16-bll data. but not
both at the same time. Note that INC HL actually Increments L. affecting H only if
L IS incremented to zero.

17)

Confusing 8- and 16-bit registers.

The Accumulator and other CPU registers are eight bits long. while the index
registers. Program Counter. Stack POinter. and register pairs are 16 bits long. You
cannot transfer the contents of a 16-blt register to an 8-bit register or vice versa.

18)

Forgetting that 16-bit numbers or addresses occupy two memory locations.

LD HL,(40H) loads Register Pair HL with the contents of memory locations 0040
and 0041. Similarly. PUSH DE stores Register Pair DE In two Stack locations. Also
remember that the Z80 stores all2-byte quantities in low-order/hlgh-order format.
For example. LD (40Hl.HL will store the contents of Register L in location 0040
and the contents of Register H in location 0041.

19)

Confusing the Stack and the Stack Pointer.

DEC. INC. and LD affect the Stack POinter. not the contents of the Stack. PUSH
and POP transfer data to or from the Stack. Remember that CALL. RET. RETI.
RETN. and RST also use the Stack to save or restore the Program Counter. The
response to an interrupt always involves saving the old Program Counter in the
Stack even if no explicit instruction IS obtained externally (as In responding to NMI
or to INT In Interrupt modes 1 or 2). Note that such instructions as EX (SPl.HL do
not affect the Stack POinter: they exchange the top two memory locations in the
Stack With the contents of a register pair or Index register. but leave the Stack
length unchanged.

20)

Forgetting to initialize the Stack Pointer.

Remember that you must place the proper memory address Into the Stack POinter
before calling any subroutines or performing any Stack operations.

21)

Changing a register or memory location before using it.

Remember that LD changes the contents of the destination (but not the source).
Be careful of instructions that Implicitly use certain registers - for example.
DJNZ decrements Register B: LDI. LDIR. LDD. LDDR. CPI. CPIR. CPO. and CPDR
all decrement the Byte Counter In Register Pair BC and Increment or decrement
Register Pair HL. LDI. LDIR. LDD. and LDDR also increment or decrement Register
Pair DE. IN!. INIR. IND. INDR. OUTI. OUTIR. aUTO. and OTOR all decrement
Register B and Increment or decrement Register Pair HL.

22)

Forgetting to transfer control past sections of the program that should not be
executed in particular situations.
Remember that the computer will proceed sequentially through the program
memory unless specifically ordered to do otherWise.

14-13

Interrupt-driven programs are particularly difficult to debug.

DEBUGGING
since errors may occur randomly. If. for example. the program
INTERRUPTenables the Interrupts a few instructions too earlv. an error will ocDRIVEN
cur only if an interrupt is received while the program is executing
PROGRAMS
those few instructions. In fact you can usually assume that randomly occurring errors are caused by the interrupt system. 2 Typical errors in interrupt-driven programs are:
1)

Forgetting to re-enable interrupts after accepting one and servicing it.

The processor disables the interrupt system automatically on RESET or on acceptIng an Interrupt. Be sure that no possible sequences fail to re-enable the Interrupt
system. Remember that. in addition to re-enabling interrupts. the program often
has to perform some action to cause the interrupting Signal to be reset. If this IS not
done. It will appear as if the interrupting device is constantly requesting service.

2)

Using the Accumulator before saving it; I.e" PUSH AF must precede any Input
or output operations thal,involve the Accumulator.

3)

Forgetting to save and restore the Accumulator and flags (Register Pair AFI.

4)

Restoring registers in the wrong order.

If the order In which thev were saved was:
PUSH
PUSH
PUSH
PUSH

AF
BC
DE
HL

the order of restoration should be:

POP
POP
POP
POP

HL
DE
BC
AF

5)

Enabling interrupts before establishing all the necessary conditions such as

priority. flags. PIO and SIO configurations. pOinters. counters. etc.
A checklist can aid here.

6)

Leaving results in registers and destroying them in the restoration process.

As noted earlier. registers should not be used to pass Information between the
regular program and the Interrupt service routines.

7)

Forgetting that RST (and NMII leaves an address in the Stack whether you
use it or not.
You may have to re-Inltlalize or update the Stack POinter.

8)

Not disabling the interrupt during multi-word transfers or instruction sequences.

Watch particularly for situations where the Interrupt service routine may use the
same memory locations that the program IS uSing.

Hopefully. these lists will at least give you some ideas as to where to look for errors. Unfortunately. even the most systematic debugging can still leave some
truly puzzling problems. particularly when interrupts are involved.3

14-14

Data ={40l

Yes

Result = (SSEG
+ Data)

(41)

Result =0

= Result

Figure 14-5. Flowchart of Decimal to Seven-Segment Conversion

14-15

Debugging Example 1: Decimal to Seven-Segment Conversion

The program converts a decimal number In memory location 0040
to a seven-segment code In memory location 0041 It blanks the
display if memory location 0040 does not contain a decimal number

DEBUGGING
A CODE
CONVERSION
PROGRAM

Initial Program (from flowchart In Figure 14-5):

DONE.
SSEG.

LD
CP
JR
LD
LD
ADD
LD
LD
HALT
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB

A,40H

9
C,DONE
HL.{SSEG)
D,A
HLDE
A,(HL)
(41H),A

GET DATA
:IS DATA A DECIMAL DIGIT?
:NO, KEEP ERROR CODE
:GET BASE ADDRESS OF 7-SEGMENT TABLE
:FIND ELEMENT BY INDEXING
:GET 7-SEGMENT CODE FROM TABLE
:SAVE 7-SEGMENT CODE OR ERROR CODE

3FH

06H
5BH
4FH

66H
6DH
7DH
07H
7DH
6FH

USing the checklist procedure. we were able to find the follOWing errors:
1)

The block that cleared Result had been omitted.

2)

The conditional branch was Incorrect.

For example. if the data IS zero. CP 9 sets the Carry, since 0 < 9. However, the lump on
the opposite condition (;.e" JR NC.DONEl still did not produce the correct result. Now
the program handles the equality case Incorrectly since. if the data IS 9. CP 9 clears the
Carry and causes a lump. The correct version IS:
CP
JR

10
NC.DONE

;IS DATA A DECIMAL DIGIT?

NO, KEEP ERROR CODE

14-16

Second Program:

DONE:
SSEG.

LD
LD
CP
JR
LD
LD
ADD
LD
LD
HALT
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB

B.O
AAOH

10
NC.DONE
HL.(SSEG)
D.A
HL.DE
A.(HLI
(41Hl.A

:GET ERROR CODE TO BLANK DISPLAY

:GET DATA
:IS DATA A DECIMAL DIGIT?
:NO. KEEP ERROR CODE
:GET BASE ADDRESS OF 7-SEGMENT TABLE
:FIND ELEMENT BY INDEXING
:GET 7-SEGMENT CODE FROM TABLE
:SAVE 7-SEGMENT CODE OR ERROR CODE

3FH
06H
5BH
4FH
66H
6DH
7DH
07H
7DH
6FH

This version was hand checked successfully.

Since the program was simple. the next stage was to sing fe-step through It with read
data. The data selected for the tnals was:

o
9
10
6B (hex)

(the smallest numbed

(the largest numbed
(a border case)
(random)

The first tnal was with zero In location 0040 (hex). The first error was obvIOus - LD
AAOH loaded the number 40 Into A. not the contents of memory location 0040. The
correct instruction was LD A(40Hl. After this correction was made. the program moved
along with no apparent errors until It tried to execute the LD A.(HLI instruction.
The contents of the Address Bus dunng the data fetch was 0647. an address that did
not even eXist In the microcomputer. Clearly. something had gone wrong.
It was now time for some more hand-checking Since we knew that JR NC.DONE was
correct. the error was beyond that Instruction but before LD A(HLI. A haf)d check
showed:
1)

LD HL. (SSEG) places 3F (hex) Into Land 06 (hex) Into H.

ThiS is clearly wrong. We want LD. HL.SSEG. not LD HL.{SSEG). That IS. we want
the address SSEG. not the contents of that address. to be loaded Into Register Pair

HL
2)

LD D.A places 0 Into Register D.

ThiS IS wrong - the data should be placed Into E. since we want to add It to the
least significant bits of the table address. In fact. an instruction should clear
Register D. since the erroneous program was not initializing or changing the other
half of Register Pair DE at all.

14-17

Third Program:
LD
LD
CP

JR

DONE.
SSEG.

LD
LD
LD
ADD
LD
LD
HALT
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB

B,O
A,(40HI
10
NCDONE
HL,SSEG

:GET ERROR CODE TO BLANK DISPLAY

:GET DATA
:IS DATA A DECIMAL DIGIT?
:NO, KEEP ERROR CODE
:GET BASE ADDRESS OF 7-SEGMENT TABLE

E.A
0,0
HL,DE
A,(HU
(41HLA

:USE DATA AS 16-BIT INDEX

:FIND ELEMENT BY INDEXING
:GET 7-SEGMENT CODE FROM TABLE
:SAVE 7-SEGMENT CODE OR ERROR CODE

3FH
06H
5BH
4FH
66H
6DH
7DH
07H
7DH
6FH

This program produced the following results:

Data

Result

00
09

3F
6F

OA

OA

68

68

The program was not clearing the reSUlt if the data was Invalid, I.e., greater than 9. The
program never used the blank code In Register B. Since the program was Simple, It
could be tested for all the decimal digits. The results were:
Data

Result

3F
06
5B
4F

2
3
4
5
6
7
8
9

69
6D
7D
07
70
6F

Note that the result lor number 8 IS wrong - It should be 7F Since everything else IS
correct the error IS almost surely In the table. In fact. entry 8 In the table had been
mlscopled.

14-18

The final program is:

DECIMAL TO 7-SEGMENT CONVERSION
LD
LD
CP

JR

DONE:

SSEG.

LD
LD
LD
ADD
LD
LD
LD
HALT
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB

B.O
A.(40HI
10
NC.DONE
HL.SSEG
E.A
D.O
HL.DE
B.(HU
A.B
(41HI.A

GET ERROR CODE TO BLANK DISPLAY

:GET DATA
.IS DATA A DECIMAL DIGIT?
:NO. KEEP ERROR CODE
:GET BASE ADDRESS OF 7-SEGMENT TABLE
:USE DATA AS 16-BIT INDEX
:FIND ELEMENT BY INDEXING
GET 7-SEGMENT CODE FROM TABLE
:SAVE 7-SEGMENT CODE OR ERROR CODE

3FH
06H
5BH
4FH
66H
6DH
7DH
07H
7FH
6FH

The errors encountered in this program are typical of the ones that
language programmers should anticipate. They include:
1)

Failing to initialize registers or memory locations.

2)

Inverting the logic on conditional branches.

zao assembly

3)

Branching Incorrectly In the case In which the operands are equal.

4)
51

Confusing Immediate and direct addressing. I.e. data and addresses.

Failing to distinguish between 8-blt data and 16-blt addresses.

61

Branching to the wrong place so that one path through the program

7)

COPYing lists of numbers (or instructions) Incorrectly.

IS

Incorrect

Note that straightforward instructions like ADD. SUB. AND. etc. seldom produce any
problems. One particularly annoYing error that you should watch for IS reversing the
operands on LD instructIOns. Many of these errors can be eliminated through the use oi
a low-level system programming language like PLZ/ASM4

14-19

Interchange flag = 1
Count = U,"gt~
of Array
= Start
of Array

Pointer

Interchange (Pointed
(Pointer + 1)
Interchenge fleg = 0

POinter = Pointer

= Count -

Count

Figure 14-6. Flowchart of Sort Program

14-20

Debugging Example 2: Sort into Decreasing Order

The program sorts an array of unsigned 8-blt binary numbers Into
decreasing order. The array begins In memory location 0041 and
ItS length IS In memory location 0040.

DEBUGGING
A SORT

PROGRAM

Initial Program (from flowchart In Figure 14-6):

PASS1

CNT

LD
LD
LD
LD
LD
INC
CP
JR
LD
INC
DJNZ
DEC
JR
HALT

C.O
A.(40H)
C.A
HL,41 H
A.(HL)
HL
(HL)
C.CNT
(HL).A
HL
PASS1

:CLEAR INTERCHANGE FLAG

:COUNT = LENGTH OF ARRAY

:WAS INTERCHANGE FLAG SET?

YES. DO ANOTHER PASS

NZ.PASS1

:POINT TO START OF ARRAY

:GET ELEMENT FROM ARRAY
:IS IT LESS THAN NEXT ELEMENT?
:NO. NO INTERCHANGE NECESSARY
YES. INTERCHANGE ELEMENTS

The hand check shows that all the blocks In the flowchart have been Implemented In
the program and that all the registers have been initialized. The conditional branches
must be examined carefully. The instruction JR C.CNT must force a branch if the new
value IS less than or equal to the old value. Note that the equality case must not result in
an interchange. since this will create an endless loop with the two equal elements
being sWitched back and forth.
Try an example:
(0040) = 30
(0041) = 37
CP (HL) results In the calculatIOn of 30-37 The Carry IS set to one. This example
should result In an interchange but does not.
JR NC.CNT will proVide the proper branch In this case. If the two numbers are equal.
the comparison will clear the Carry and JR NC.CNT is again correct.
How about JR NZ.SORT at the end of the program? If there are any elements out of
order. the interchange flag will be one. so the branch IS wrong. It should be JR Z.SORT.
Now let's hand check the first Iteration of the program. The initialization results In the
following values:
A
B

C
HL

COUNT
COUNT

0
0041

The effects of the loop instructions are:

CNT

LD
INC
CP
JR
LD
INC
DJNZ

A.(HL)
HL
(HL)
NC.CNT
(HU.A
HL
PASS1

:A = (0041)
:HL = 0042
:(0041 )-(0042)
:(0042) = (0041)
:HL = 0043
;B = COUNT-1

14-21

Note that we have already checked the Conditional Jump instructions. Clearly the logic
is Incorrect. If the first two numbers are out of order, the results after the first Iteration
should be:
(0041)
(0042)
HL
B

OLD (0042)
OLD (0041)
0042
COUNT-1

(0041)
(0042)
HL

UNCHANGED
OLD (0041)
0043
COUNT-1

Instead, they are:

The error in HL IS easy to correct. The second INC HL is unnecessary and should be
omitted. The Interchange requires a bit more care and a temporary register, Le.,
LD
LD
DEC
LD
INC

D,(HL)
(HL),A
HL
(HL),D
HL

An Interchange always requires a temporary storage place in which one number can be
saved while the other one IS being transferred.

14-22

All of these changes require a new copy of the program. i.e.

PASS1:

CNT

LD
LD
LD
LD
LD
INC
CP
JR
LD
LD
DEC
LD
INC
DJNZ
DEC
JR
HALT

:CLEAR INTERCHANGE FLAG

;COUNT = LENGTH OF ARRAY

C.O
A (40H)
C.A
HL,41 H
A(HL)
HL
(HU
NC.CNT
D.(HU
(HL).A
HL
(HL).D
HL
PASS1

;POINT TO START OF ARRAY

:GET ELEMENT FROM ARRAY
:IS IT LESS THAN NEXT ELEMENT?
;NO. NO INTERCHANGE NECESSARY
:YES. INTERCHANGE ELEMENTS

C
NZ.PASS1

:WAS INTERCHANGE FLAG SET?

:YES. DO ANOTHER PASS

How about the last iteration? Let's say that there are three elements:
(0040)
(0041)
(0042)
(0043)

03
02
04
06

Each time through. the program Increments Register Pair HL by one. So. at the start of
the third Iteration.
(HL)

= 0041

+ 2 = 0043

The effects of the loop instructions are:

LD
INC
CP

A.(HL)
HL
(HL)

:A = (0043)
;HL =0044
;(0043)-(0044)

ThiS IS incorrect the program has tned to move beyond the end of the data. The previous Iteration should. In fact. have been the last one. since the number of pairs IS one
less than the number of elements. The correction IS to reduce the number of Iterations
by one: thiS can be accomplished by plaCing DEC B after LD A.(40H).

How about the trivial cases? What happens if the array contains no elements at
all. or only one element? The answer is that the program does not work correctly
and may change a whole block of data improperly and without any warning (try
itO. The corrections to handle the trivial cases are simple but essential; the cost
is only a few bytes of memory to avoid problems that could be very difficult to
solve later.

14-23

The new program is:

PASS1

CNT

LD
LD
CP
JR
LD
DEC
LD
LD
INC
CP
JR
LD
LD
DEC
LD
INC
DJNZ
DEC
JR
HALT

C.O
A.(40H)

2
C.DONE
B.A
B
HL.41 H
A.(HU
HL
(HU
NC.CNT
D.(HU
(HU.A
HL
(HU.D
HL
PASS1

C
NZ.PASS1

CLEAR INTERCHANGE FLAG

COUNT = LENGTH OF ARRAY
DOES ARRAY HAVE 2 OR MORE ELEMENTS?
;NO, NO ACTION NECESSARY
:NUMBER OF PAIRS = COUNT-1
;POINT TO START OF ARRAY
:GET ELEMENT FROM ARRAY
;IS IT LESS THAN NEXT ELEMENT?
:NO, NO INTERCHANGE NECESSARY
;YES. INTERCHANGE ELEMENTS

:WAS INTERCHANGE FLAG SET?

:YES, DO ANOTHER PASS

Now It'S time to check the program on the computer or on the Simulator. A simple set of
data IS:
(0040)

02

(0041)

00

(0042)

01

ThiS set consists of two elements In the wrong order. The program should take two
passes. The first pass should rearrange the elements. producing:
(0041)
(0042) =
C =

01
00
01

The second pass should complete the operation and produce:

C = 00
ThiS program IS rather long for single stepping, so we'll use breakpoints Instead. Each
breakpoint will halt the computer and pnnt the contents of all the registers. The breakPOints will come:
1)

After LD HL,41 H to check the initialization.

2)

After CP (HU to check the companson.

3)

After the second INC HL (i.e.. lust before the label CNT) to check the Interchange.

4)

After DEC C to check the completion of a pass through the array. The contents oj
the registers after the first breakpoint were:
Register

Contents

A
B

01

C
H

00
00

41

02

These are all correct. so the program is performing the initialization correctly In thiS
case.

14-24

The results at the second breakpoint were:

Register

Contents

A
B
C

00
01
00
00

H
L
CARRY

42
1

These results are also correct. The results at the third breakpOint were:
Register

Contents

A
B

00
01
00
01
00

42

o
Checking memory showed:

(0041) = 01
(0042) = 00
The results at the fourth breakpoint were:
Register

Contents

00

01

00
01

00

42

Here. Register C does not contain the correct value - it should have been set to one to
Indicate that an Interchange had occurred. In fact. a look at the program shows that no
instruction ever changes C to mark the Interchange. The correction IS to place the instruction LD C.l after JR NC.CNT
Now the procedure IS to load Register C With the correct value and continue. The second Iteration of the second breakpOint gives:
Register

Contents

A
B

00
00
00
00
43
1

H
L
CARRY

Clearly the program has proceeded Incorrectly Without relnltlalizing the registers (particularly HL). The conditional Jump that depends on the Interchange flag should transfer
control all the way back to the start of the program. not to the label PASS1.

14-25

The final version of the program is:

SORT

PASS1

LD
LD
CP
JR
LD
DEC
LD
LD
INC
CP
JR
LD
LD
LD
DEC
LD

INC
CNT

DJNZ
DEC
JR
HALT

C.O
A.(40H)
2
C.DONE
B.A
B
HL,41 H
A.(HU
HL
(HU
NC.CNT

c:,

D.(HU
(HU.A
HL
(HU.D
HL
PASS1
C
NZ.SORT

;CLEAR INTERCHANGE FLAG

;COUNT = LENGTH OF ARRAY
;DOES ARRAY HAVE 2 OR MORE ELEMENTS?
;NO. NO ACTION NECESSARY
;NUMBER OF PAIRS = COUNT-1
;POINT TO START OF ARRAY
;GET ELEMENT FROM ARRAY
;IS IT LESS THAN NEXT ELEMENT?
;NO. NO INTERCHANGE NECESSARY
;YES. SET INTERCHANGE FLAG
;INTERCHANGE ELEMENTS

;WAS INTERCHANGE FLAG SET?

;YES. DO ANOTHER PASS

Clearly we cannot check all the possible Input values for this program. Two other simple
sets of data for debugging purposes are;
1)

Two equal elements

(0040)
(0041)
(0042)

2)

02
00
00

Two elements already In decreasing order

(0040)
(0041)
(0042)

02
01
00

14-26

INTRODUCTION TO TESTING
Program testing is closely related to program debugging.
Surely some of the test cases will be the same as the test
data used for debugging, such as:

USING TEST
CASES FROM
DEBUGGING

Trivial cases such as no data or a single element

Special cases that the program singles out for some reason
Simple examples that exercise particular parts of the program
In the case of the decimal to seven-segment conversion program, these cases
cover all the possible situations. The test data consists of:
The numbers 0 through 9
The boundary case 10
The random case 68
The program does not distinguish any other cases. Here debugging and testing are
virtually the same.

In the sorting program, the problem is more difficult. The number of elements could
range from 0 to 255. and each of the elements could lie anywhere In that range. The
number of possible cases IS therefore enormous. Furthermore. the program IS
moderately complex. How do we select test data that will give us a degree of confidence in that program? Here testing requires some design decisions. The testing
problem is particularly difficult if the program depends on sequences of real-time data.
How do we select the data. generate It. and present It to the microcomputer in a
realistic manner?
Most of the tools mentioned earlier for debugging are helpful
in testing also. Logic or microprocessor analyzers can help
check the hardware; simulators can help check the software.
Other tools can also be of assistance, e.g.,
1)
2)
3)
4)

5)
6)
7)
8)

I/O simulations that can simulate a variety of devices from a single Input and a
single output device.
In-circuit emulators that allow you to attach the prototype to a development
system or control panel and test It.
ROM simulators that have the flexibility of a RAM but the timing of the particular
ROM or PROM that will be used in the final system.
Real-time operating systems that can provide Inputs or interrupts at specific
times (or perhaps randomly) and mark the occurrence of outputs. Real-time breakPOints and traces may also be included.
Emulations (often on micro programmable computers) that may provide real-time
execution speed and programmable 1/0. 5
Interlaces that allow another computer to control the I/O system and test the
microcomputer program.
Testing programs that check each branch In a program for logical errors.
Test generation programs that can generate random data or other distributions.

Formal testing theorems eXIst. but they are usually applicable only to very short programs.

You must be careful that the test equipment does not invalidate the test by
modifying the environment. Often, test equipment may buffer, latch, or condition
input and output signals. The actual system may not do this, and may therefore
behave quite differently.

14-27

Furthermore. extra software in the test environment may use some of the memory space or part of the interrupt system. It may also provide error recovery and
other features that will not exist in the final system. A software test bed must be
lust as realistic as a hardware test bed. since software failure can be lust as Critical as
hardware failure.
Emulations and simulations are. of course. never precise. They are usually adequate for checking logic. but can seldom help test the interface or the timing. On
the other hand. real-time test equipment does not provide much of an overview of
the program logic and may affect the interfacing and timing.

SELECTING TEST DATA

Very few real programs can be checked for all cases. The designer must choose a
sample set that in some sense describes the entire range of possibilities.
Testing should. of course. be part of the total development proSTRUCTURED
cedure. Top-down deSign and structured programming proVide for
TESTING
testing as part of the deSign. ThiS IS called structured testlng.6
Each module within a structured program should be checked separately. Testing. as
well as design. should be modular. structured. and top-down.

But that leaves the question of selecting test data for a

module. The designer must first list all special cases that a
program recognizes. These may Include:

TESTING
SPECIAL
CASES

TriVial cases
Equality cases
Special situations
The test data shou Id Include all of these.

You must next identify each class of data that statements

within the program may distinguish. These may Include:
Positive or negative numbers
Numbers above or below a particular threshold

FORMING
CLASSES
OF DATA

Data that does or does not Include a particular sequence or character

Data that IS or IS not present at a particular lime
If the modules are short. the total number of classes should still be small even though
each diVISion IS multiplicative: I.e.. two two-way diVISions result In four data classes.

You must now separate the classes according to whether the

SELECTING
program produces a different result for each entry in the class
DATA FROM
(as in a table) or produces the same result for each entry (such
CLASSES
as a warning that a parameter is above a thresholdl. In the discrete case. one may Include each element if the total number IS small or sample if the
number IS large. The sample should include all boundary cases and at least one case
selected randomly. Random number tables are available In books. and random number
generators are part of most computer facilities.
You must be careful of distinctions that may not be obvious. For example. an 8-bit
microprocessor will regard an 8-bit unsigned number greater than 127 as negative; the programmer must consider this when using conditional branches that
depend on the Sign flag. You must also watch for instructions that do not affect
flags, overflow in signed arithmetic, and the distinctions between address-length
116-bitl quantities and data-length 18-bit) quantities.

14-28

Testing Example 1: Sort Program

The special cases here are obvIous:

TESTING
A SORT
PROGRAM

No elements in the array

One element. magnitude may be selected randomly

The other special case to be considered IS one in which elements are equal.
There may be some problem here with signs and data length. Note that the array Itself
must contain fewer than 256 elements. The use of the instruction LD C.l or SET 1.C
rather than DEC C to clear the interchange flag means that there will be no difficulty if
the number of elements or interchanges exceeds 128.
We could check the effects of sign by picking half the regular test cases with numbers
of elements between 128 and 255 and half between 2 and 127 All magnitudes should
be chosen randomly so as to avoid unconscious bias as much as possible.

Testing Example 2: Self-Checking Numbers

.....

(see Chapter 8)

Here we will presume that a prior validity check has ensured that
the number has the right length and consists of valid digits. Since
the program makes no other distinctions. test data should be
selected randomly. Here a random number table or random number generator will prove Ideal: the range of the random numbers IS 0

~~--..,

TESTING AN
ARITHMETIC
PROGRAM

to 9.

TESTING PRECAUTIONS
The designer can simplify the testing stage by designing programs sensibly. You should use the following rules:
1)

Try to eliminate trivial cases as early as possible without introducing unnecessary distinctions.

21

Minimize the number of special cases. Each special case means additional testing
and debugging time.
Consider performing validity or error checks on the data prior to processing.

3)
4)

Be careful of inadvertent and unnecessary distinctions. partlcularlv In handling

signed numbers or uSing operations that refer to signed numbers.

51

Check boundary cases by hand. These are often a source of errors. Be sure that the
problem definitIOn specifies what IS to happen In these cases.
Make the program as general as reasonably possible. Each distinction and separate
routine Increases the required testing.

6)
71

DiVide the program and deSign the modules so that the testing can proceed In
steps In cOnlunction With the other stages of software development 7

14-29

CONCLUSIONS
Debugging and testing are the stepchildren of the software development process.
Most projects leave far too little time for them and most textbooks neglect them.
But designers and managers often find that these stages are the most expensive
and time-consuming. Progress may be very difficult to measure or produce.
Debugging and testing microprocessor software is particularly difficult because
the powerful hardware and software tools that can be used on larger computers
are seldom available for microcomputers.
The designer should plan debugging and testing carefully. We recommend the
following procedure:
1)

2)
3)
4)

5)

6)
7)
8)

Try to write programs that can easily be debugged and tested. Modular programming. structured programming. and top-down design are useful techniques.
Prepare a debugging and testing plan as part of the program design. Decide
early what data you must generate and what equipment you will need.
Debug and test each module as part of the top-down design process.
Debug each module's logic systematically. Use checklists. breakpoints. and
the single-step mode. If the program logic is complex. consider using the software simulator.
Check each module's timing systematically if this is a problem. An
oscilloscope can solve many problems if you plan the test properly. If the timing is complex. consider using a logic or microprocessor analyzer.
Be sure that the test data is a representative sample. Watch for any classes of
data that the program may distinguish. Include all special and trivial cases.
If the program handles each element differently or the number of cases is
large. select the test data randomly.S
Record all test results as part of the documentation. If problems occur. you
will not have to repeat test cases that have already been checked.

14-30

REfERENCES
1.

For more information about logic analyzers. see:

R. L. Down. "Understanding Logic Analyzers:' Computer Design. June 1977. pp.
188-191.
-W. A. Farnbach. "Bring up Your (J.P." Electronic Design. July 10. 1976. pp. 80-85.
B. Farly. "Logic Analyzers Aren't All Alike:' Electronic Design. Feb. 1. 1978. pp.
70-76.
-K. Pines. "What Do Logic Analyzers Do?" Digital Design. September 1977, pp.
55-77
---N. A. Robin. "The logic Analyzer: A Computer Troubleshooting TooL" Computer
Design. March 1976. pp. 89-96.
S. Runyon. "Focus on Logic and (J.P Analyzers:' Electronic Design. February 1.
1977, pp. 40-50.
-A. Santoni. "The Latest Logic Analyzers Offer More Functions and Less Cost:'
Electronic Design. Feb. 1. 1978. pp. 26-32.

2.

See W. J. Weller. Assembly Level Programming for Small Computers. Lexington

Books. Lexington. Mass.. 1975.

3.

Some guidelines for debugging Interrupt problems are given In R. L. Baldrige. "Interrupts Add Power. Complexity to Microcomputer System Design:' EDN. August
5. 1977, pp. 67-73.

4.

See C. Bass. "PLZ: A Family of System Programming Languages for

Microprocessors:' Computer. March 1978. pp. 34-39.

5.

See. for example. H. R. BUrris. "Time-Scaled Emulations of the 8080

Microprocessor:' Proceedings of the 1977 National Computer Conference. pp.
937-946.

6.

See D. A. Walsh. "Structured Testing:' Datamation. July 1977, pp. 111-118.

Testing (and debugging) are also discussed in R. A. DeMilio et al.. "Hints on Test
Data Selection: Help for the Practicing Programmer." Computer. April 1978. pp.
34-41 and in W. F, Dalton. "Design Microcomputer Software:' Electronics. January
19.1978. pp. 97-101.

8.

Random numbers and their generation are discussed in 1. G. Lewis. Distribution

Sampling for Computer Simulation. Lexington Books. Lexington. Mass.. 1975 and
In R. A. Mueller. et al.. "A Random Number Generator for Microprocessors:' Simulation. April 1977, pp. 123-127
---

14-31

Chapter 15
DOCUMENTATION AND REDESIGN
The working program is not the only requirement of software development. Adequate documentation is also an important part of a software product. Not only
does documentation help the designer in the testing and debugging stages. it is
also essential for later use and extension of the program. A poorly documented
program will be difficult to maintain. use. or extend.
Occasionally. a program uses too much memory or executes too slowly. The
designer must then improve it. This stage is called redesign. and requires that you
concentrate on the parts of the program that can yield the most improvement.

SELF-DOCUMENTING PROGRAMS
Although no program is ever completely self-documenting. some of the rules that we mentioned earlier can help.
These include:

RULES FOR
SELF-DOCUMENTING
PROGRAMS

Clear. Simple structure with as few transfers of control

(jumps) as possible
Use of meaningful names and labels
Use of names for I/O deVices. parameters. numerical factors. etc.
EmphasIs on simplicity rather than on minor savings In memory usage. execution
lime. or tYPing
For example. the following program sends a string of characters to a teletypewriter:

W:

LD
LD
LD
LD
OUT
CALL
INC
DJNZ
HALT

A.(2000H)
B.A
HL.l000H
A.(HU
(61.A

XXX
HL

Even without comments we can Improve the program. as follows:

MESSG
COUNT
TTYSIO

OUTCH:

EQU
EQU
EQU
LD
LD
LD
LD
OUT
CALL
INC
DJNZ
HALT

1000H
2000H

6
A.(COUNT)
B.A
HL.MESSG
A.(HU
(TTYSIOl.A
BITDLY
HL
OUTCH

15-1

--~~---~~ - - - -

Surely this program IS easier to understand than the earlier version~ Even without
further documentation. you could probably guess at the function of the program and
the meanings of most of the variables. Other documentation techniques cannot
substitute for self-documentation.

Some further notes on choosing names:

CHOOSING

1)

USEFUL

2)

3)
4)

Use the obvious name when It is available, like TTY or CRT

NAMES
for output devices, START or RESET for addresses. DELAY or
SORT for subroutines. COUNT or LENGTH for data.
Avoid acronyms like S168A for ..ORT llHilT ,6RRAY These seldom mean anything to anybody.
Use full words or close to full words when possible, like DONE. PRINT. SEND. etc.
Keep the names as distinct as possible.

COMMENTS
The most obvious form of additional documentation is the comment. However,
few programs (even those used as examples in books!. have effective comments.
You should consider the following guidelines for good comment..s;..
--,
1)

Don't repeat the meaning of the instruction code. Rather.

explain the purpose of the instruction in the program. Comments like
;8 = 8-1
DEC 8
add nothing to documentation. Rather. use
DEC 8
;L1NE NUM8ER = LINE NUM8ER-1

COMMENTING
GUIDELINES

Remember that you know what the operation codes mean and anyone else can
look them up in the manual. The important point is to explain what task the
program is performing.
21

Make the comments as clear as possible. Do not use abbreviations or acronyms

unless they are well-known (like ASCII. PIO. or UART) or standard (like no for number, ms for millisecond. etc'!' Avoid comments like
DEC 8
:LN = LN-1
or
;DEC LN 8Y 1
DEC 8
The extra typing simply is not all that expensive.

3)

Comment every important or obscure point. 8e particularly careful to mark

operations that may not have obvIous functions, such as
AND 110111118
;TURN TAPE READER 81T OFF
or
ADD HL.DE
;INDEX GRAY CODE TA8LE
Clearly. I/O operations often require extensive comments. If you're not exactly
sure of what an instruction does, or if you have to think about It. add a clarifYing
comment. The comment will save YOU time later and will be helpful in documentation.

15-2

4)

Don't comment the obvious. A comment on each line simply makes It difficult to
find the Important points. Standard sequences like
INC
HL
DJNZ
SEARCH
need not be marked unless you're dOing something special. One comment will
often suffice for several lines. as In
RRCA
;SWAP DIGITS
RRCA
RRCA
RRCA
LD
LD
LD

A.C
C.B
B.A

;EXCHANGE MOST SIGNIFICANT. LEAST

SIGNIFICANT BYTES

5)

Place comments on the lines to which they refer or at the start of a sequence.

6)

Keep your comments up-to-date. If you change the program. change the comments.

7)

Use standard forms and terms In commenting. Don't worry about repetitiveness.
Varied names for the same things are confusing. even if the vanatlons are lust
COUNT and COUNTER. START and BEGIN. DISPLAY and LEOS. or PANEL and
SWITCHES.
There's no real gain In not being consistent. The vanatlons may seem obvious to
you now. but may not be clear later; others will get confused from the very beginning.

8)

Make comments mingled with instructions brief. Leave a complete explanation

to header comments and other documentation. Otherwise. the program gets lost
In the comments and YOU may have a hard time even finding it.

9)

Keep improving your comments. If you come to one that you can't read or understand. take the time to change It. If you find that the listing is getting crowded.
add some blank lines. The comments won't Improve themselves; In fact. they will
lust become worse as you leave the task behind and forget exactly what you did.

10)

Before every major section, subsection, or subroutine, insert a number of

comments describing the functions of the code that follows. Care should be
taken to describe all Inputs. outputs. and side effects. as well as the algonthm
employed.

11)

It IS good practice when modifying working programs to use comments to indicate the date, author, and type of modification made.

Remember, comments are important. Good ones will save you time and effort. Put
some work into comments and try to make them as effective as possible.

15-3

Commenting Example 1: Multiple-Precision

Addition

COMMENTING
EXAMPLES

The basic program is:

ADDWD:

LD
LD
LD
LD
AND
LD
ADC
LD
INC
INC
DJNZ
HALT

A,(30H)
B,A
HL,41H
DE,51H

A
A,(DE)
A,(HL)
(HL),A
DE
HL
ADDWD

First. comment the important POints. These are typically Inltializations, data fetches,
and processing operations. Don't bother with standard sequences like updating pointers and counters. Remember that names are clearer than numbers, so use them freely.
The new version of the program IS:
:MULTIPRECISION ADDITION
;THIS PROGRAM PERFORMS MULTI-BYTE ADDITION
;INPUTS:

:OUTPUTS:
LENGTH
NUMB1
NUMB2

ADDWD:

LOCATION 30H = LENGTH OF NUMBERS (IN BYTES)

LOCATIONS 41 H-50H = FIRST ADDEND IN LSB-MSB ORDER
LOCATIONS 51 H-60H = SECOND ADDEND
LOCATIONS 41 H-51 H = SUM
EQU
EQU
EQU
LDA
LD
LD
LD
AND
LD
ADC
LD
INC
INC
DJNZ
HALT

30H
41H
51H
LENGTH
B,A
HL.NUMB1
DE,NUMB2

;COUNT = LENGTH OF NUMBERS (IN BYTES)

;START AT LSB'S OF 1ST NUMBER
;START AT LSB'S OF 2ND NUMBER

A
A,(DE)
A,(HU
(HL),A
DE
HL
ADDWD

;GET 8 BITS OF 2ND NUMBER

;ADD 8 BITS OF 1ST NUMBER
;STORE RESULT IN 1ST NUMBER

Second, look for any instructions that might not have obvious
functions and mark them. Here, the purpose of AND A IS to clear
the Carry the first time through.

QUESTIONS
FOR

COMMENTING

Third, ask yourself whether the comments tell you what you would
need to know if you wanted to use the program, e.g.:
1)

Where is the program entered? Are there alternative entry POints?

2)

What parameters are necessary? How and in what form must they be supplied?

15-4

3)

What operations does the program perform?

4)

From where does it get the data?

5)

Where does it store the results?

6)

What special cases does it consider?

7)

What does the program do about errors?

8)

How does It exit?

Some of the questions may not be relevant to a particular program and some of the
answers may be obvious. Make sure that you won't have to sit down and dissect the
program to figure out what the answers are. Remember that too much explanation is
lust dead wood that you will have to clear out of the way. Is there anything that you
would add to or subtract from this listing? If so, go ahead - yOU are the one who has to
feel that the commenting IS adequate and reasonable.
. MULTIP RECISION ADDITION
;THIS PROGRAM PERFORMS MULTI-BYTE ADDITION
LOCATION 30H = LENGTH OF NUMBERS (IN BYTES)
LOCATIONS 41H-50H = FIRST ADDEND IN LSB-+MSB ORDER
LOCATIONS 51 H-60H = SECOND ADDEND
OUTPUTS;LOCATIONS 41H-51 H = SUM
INPUTS;

LENGTH
NUMBl
NUMB2

ADDWD:

EQU
EQU
EQU
LOA
LD
LD
LD
AND
LD
ADC
LD
INC
INC
DJNZ
HALT

30H
41H
51H
LENGTH
B,A
HL,NUMBl
DE.NUMB2

A
A,(DE)
A(HLl
(HU,A
DE
HL
ADDWD

;LENGTH OF NUMBERS
;LSB'S OF 1ST NUMBER AND RESULT
;LSB'S OF 2ND NUMBER
;COUNT = LENGTH OF NUMBERS (IN BYTES)
;START AT LSB'S OF 1ST NUMBER
;START AT LSB'S OF 2ND NUMBER
;CLEAR CARRY TO START
;GET 8 BITS OF 2ND NUMBER
;ADD 8 BITS OF 1ST NUMBER
;STORE RESULT IN 1ST NUMBER

Commenting Example 2: Teletypewriter Output

The basic program is:

TBIT

LD
ADD
LD
OUT
RRA
SCF
CALL
DJNZ
HALT

A,(60H)
AA
B,ll
(PIODRBl.A

BITDLY
TBIT

15-5

Commenting the important POints and adding names gives:

:TELETYPEWRITER OUTPUT PROGRAM
:THIS PROGRAM PRINTS THE CONTENTS OF MEMORY LOCATION 60H TO THE
TELETYPEWRITER
INPUTS: LOCATION 60H = CHARACTER CODE
OUTPUTS: NONE
TIYPIO
NBITS
TDATA

TBIT

EQU
EOU
EQU

PIODRB
11
60H

LD
ADD
LD
OUT
RRA
SCF
CALL
DJNZ
HALT

A, (TDATA)

A.A
B.NBITS
(TTYPIO).A

BITDLY
TBIT

:NUMBER OF BITS PER CHARACTER

:ADDRESS OF CHARACTER TO BE
. TRANSMITIED
:GET DATA
:SHIFT LEFT AND FORM START BIT
:COUNT = NUMBER OF BITS PER CHARACTER
:SEND BIT TO TIY
:UPDATE FOR NEXT BIT
:FORM STOP BIT (LOGIC ONE)
:DELAY 1 BIT TIME

Note how easily we could change this program so that It would transfer a whole string
of data. starting at the address in locations DPTR and DPTR + 1 and ending with an
"03" character (ASCII ETX). Furthermore. let us make the terminal a 30 character per
second device with one stop bit (we will have to change subroutine BITDLYl. Try makIng the changes before looking at the listing.
;STRING OUTPUT PROGRAM
:THIS PROGRAM OUTPUTS A STRING TO THE TERMINAL. TRANSMISSION CEASES
WHEN AN ASCII ETX (30H) IS ENCOUNTERED
INPUTS:

LOCATIONS 60H-61 H CONTAIN ADDRESS OF

STRING TO OUTPUT
OUTPUTS: NONE
DPTR

EQU

60H

ENOCH
NBITS
TIYPIO

EQU
EQU
EQU
LD
LD
CP
JR
ADD
LD
OUT
RRA
SCF
CALL
DJNZ
INC
JR
HALT

03
11
PIODRB
HL.(DPTR)
A.(HU
ENOCH
Z.DONE
A.A
B.NBITS
(TTYPIO).A

TCHAR:

TBIT:

DONE:

BITDLY
TBIT
HL
TCHAR

;LOCATION OF OUTPUT BUFFER START

. ADDRESS
:ENDING CHARACTER = ASCII ETX
:NUMBER OF BITS PER CHARACTER
:GET STARTING ADDRESS OF STRING
:GET A CHARACTER
:IS IT ENDING CHARACTER?
:YES. DONE
;SHIFT DATA LEFT AND FORM START BIT
:COUNT = NUMBER OF BITS PER CHARACTER
:SEND BIT TO TIY
:UPDATE FOR NEXT BIT
:FORM STOP BIT (LOGIC ONE)
;DELAY 1 BIT TIME

15-6

Good comments can make it easy for you to change a program to meet new requIrements. For example. try changing the last program so that it:
Starts each message with ASCII STX (02 hex) followed by a three-digit identification
code stored In memory locations 0030 through 0032
Adds no start or stop bits
Walts 1 ms between bits
Transmits 40 characters. starting with the one located at the address In DPTR and
DPTR+1
Ends each message with two consecutive ASCII ETXs (03 hex)

FLOWCHARTS AS DOCUMENTATION
We have already described the use of flowcharts as a design tool
In Chapter 13. Flowcharts are also useful in documentation. particularly if:

HINTS FOR
USING
FLOWCHARTS

They are not so detailed as to be unreadable

Their decision points are clearlY explained and marked
They Include all branches
They correspond to the actual program listings
Flowcharts are helpful if they give you an overall picture of the program. They are not
helpful if they are just as difficult to read as an ordinary lisllng.

STRUCTURED PROGRAMS AS DOCUMENTATION

A structured program can serve as documentation for an assembly language program
if:
You describe the purpose of each section In the comments
You make it clear which statements are included In each conditional or loop structure
by uSing indentation and ending markers
You make the total structure as Simple as possible
You use a consistent. well-defined language
The structured program can help you to check the logic or improve it. Furthermore.
since the structured program IS machine-independent. It can also aid you in implementIng the same task on another computer.

MEMORY MAPS
A memory map is Simply a list of all the memory assignments in a program. The map
allows YOU to determine the amount of memory needed. the locations of data or
subroutines. and the parts of memory not allocated. The map is a handy reference for
finding storage locations and entry points and for dividing memory between different
routines or programmers. The map will also give you easy access to data and
subroutines if YOU need them In later extensions or in maintenance. Sometimes a
graphical map IS more helpful than a listing.

15-7

TYPICAL
MEMORY
MAP

A typical map would be:

Program Memory
Address

Routine

Purpose

0000-0002

RESET

0038-003A

INTRPT

0040-0265
0270-027F
0280-0290
0300-0340

MAIN
DELAY
DSPLY
KEYIN

TRANSFERS CONTROL TO MAIN PROGRAM IN LOCATION

40 HEX
TRANSFERS CONTROL TO INTERRUPT SERVICE
IN LOCATION 300 HEX
MAIN PROGRAM
DELAY PROGRAM
DISPLAY CONTROL PROGRAM
INTERRUPT CONTROL PROGRAM FOR KEY80ARD
Data Memory

1000
1001-1002
1003-1041
1042-1051
1052-105F
10EO-10FF

NKEYS
KPTR
KBFR
DBFR
TEMP
STACK

NUMBER OF KEYS
KEYBOARD BUFFER POINTER
KEYBOARD BUFFER
DISPLAY BUFFER
TEMPORARY STORAGE
RAM STACK

The map may also list additional entry POints and include a specific deSCription of the
unused parts of memory.

PARAMETER AND DEFINITION LISTS

Parameter and definition lists at the start of the program and each subroutine
make understanding and changing the program far simpler. The following rules can
help:

1) Separate RAM locations, I/O units, parameters, definitions, and memory system constants.
2)

Arrange lists alphabetically when possible, With a deSCrip-

RULES FOR
DEFINITION
LISTS

tion of each entry.

3)

Give each parameter that might change a name and include it in the lists. Such
parameters may Include timing constants, inputs or codes corresponding to particular keys or functions, control or masking patterns, starting or ending characters,
thresholds, etc.

4)

Make the memory system constants into a separate list. These constants will
Include Reset and Interrupt service addresses, the starting address of the program,
RAM areas, Stack areas, etc.

5)

Give each port used by an I/O device a name, even though devices may share
ports in the current system. The separation will make expansion or reconfiguratlon
much simpler.

15-8

A typical list of definitions will be:

TYPICAL
DEFINITION
LIST

:MEMORY SYSTEM CONSTANTS

RESET
INTRP
START
KEYIN
RAMST
STKPTR

EQU
EQU
EQU
EQU
EQU
EQU

0
3SH
40H
300H
1000H
1100H

:RESET ADDRESS
:INTERRUPT ENTRY
:START OF MAIN PROGRAM
:KEYBOARD INTERRUPT PROGRAM
:START OF DATA STORAGE
:START OF STACK

EQU
EQU
EQU
EQU

OEOH
OE1H
OEOH
OFOH

:OUTPUT PIO FOR DISPLAYS

:INPUT PIO FOR KEYBOARD
:OUTPUT PIO FOR KEYBOARD
TTY DATA PORT

. I/O UNITS
DSPLY
KBDIN
KBDOUT
TTYPIO

:RAM LOCATIONS

NKEYS
KBDPTR
KBDBFR
DSPBFR
TEMP

ORG
DEFS
DEFS
DEFS
DEFS
DEFS

RAMST
1

2
40H
lOH
14H

:NUMBER OF KEYS
:KEYBOARD BUFFER POINTER
:KEYBOARD INPUT BUFFER
:DISPLAY DATA BUFFER
TEMPORARY STORAGE

:PARAMETERS
BOUNCE
GOKEY
MSCNT
OPEN
TPULS

EQU
EQU
EQU
EQU
EQU

10
133
OFH
1

:DEBOUNCING TIME IN MS
:IDENTIFICATION OF 'GO' KEY
:COUNT FOR 1 MS DELAY
:PATTERN FOR OPEN KEYS
;PULSE LENGTH FOR DISPLAYS IN MS

EQU
EQU

OFFH
SOH

:ALL ONES PATTERN

:START CONVERSION PULSE

;DEFINITIONS
ALL1
STCON

Of course, the RAM entnes will usually not be In alphabetical order, since the designer
must order these so as to minimize the number of address changes required In the program.

15-9

LIBRARY ROUTINES
Standard documentation of subroutines will allow you to build up a library of
useful programs. The Idea is to make these programs easily accessible. A standard format will allow you or anyone else to see at a glance what the program does. The best
procedure is to make up a standard form and use It consistently. Save these programs
In a well-organized manner (for example, according to processor, language, and type of
program), and you will soon have a useful set. But remember that without organization and proper documentation. using the library may be more difficult than rewriting the program from scratch. Debugging a system requires a precise understanding
of all the effects of each subroutine.
Among the information that you will need In the standard form IS:
Purpose of the program
Processor used

STANDARD
PROGRAM
LIBRARY
FORMS

Language used
Parameters required and how they are passed to the subroutine
Results produced and how they are passed to the main program
Number of bytes of memory used
Number of clock cycles required. ThiS number may be an average or a tYPical figure,
or It may vary Widely. Actual execution time will. of course, depend on the processor
clock rate
Registers affected
Flags affected
A tYPical example
Error handling
SpeCial cases
Documented program listing
If the program IS complex, the standard library form should also Include a general
flowchart or a structured program. As we have mentioned before, a library program IS
most likely to be useful if It performs a single distinct function In a reasonably general
manner.

LIBRARY EXAMPLES
Library Example 1: Sum of Data
Purpose: The program SUM8 computes the sum of a set of 8-blt unsigned binary numbers.
Language: Z80 assembler.
Initial Conditions: Starting address of set of numbers In Register Pair HL length of set
In Accumulator.
Final Conditions: Sum In Accumulator.
Requirements:
Memory
Time

7 bytes.
13 + 26N clock cycles, where N IS the
length of the set of numbers.
Registers
A, B, H. L
All flags affected.

15-10

Typical Case: (all data in hexadecimal)

Start:
HL

0050

03

A
(0050)
(0051)
(0052)
End:

27

3E
26

88

Error Handling: Program Ignores all carries. Carry bit reflects only the last operation.
Initial contents of Accumulator must be 1 or more.
Listing:
:SUM OF 8-BIT DATA
SUM8:
ADD8:

LD
SUB
ADD
INC
DJNZ
RET

B.A
A
A.(HL)
HL
ADD8

:COUNT = LENGTH OF DATA BLOCK

:SUM = ZERO
:SUM = SUM + DATA ENTRY

Library Example 2: Decimal-to-Seven-Segment Conversion

Purpose: The program SEVEN converts a decimal number to a seven-segment display
code.
Language: Z80 assembler.
Initial Conditions: Data in Accumulator.
Final Conditions: Seven-segment code in Accumulator.
Requirements:
Memory

26 bytes. including the seven-segment table (10 entnes!.

Time
- 74 clock cycles if the data IS valid. 40 if it is not.
A. B. D. E. H. L.
Registers
All flags affected.
Input data in Accumulator is destroyed.

Typical Case: (data

In

hexadecimal)

Start:

05

66

End:

Error Handling: Program returns zero in the Accumulator if data IS not a deCimal digit.

15-11

Listing:
:DECIMAL TO SEVEN-SEGMENT CONVERSION
SEVEN:

LD
CP

B.O
10
NC.DONE
L.A
H.O
DE.SSEG
HL.DE
B.(HU
A.B

JR

DONE:
SSEG:

LD
LD
LD
ADD
LD
LD
RET
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB

:GET ERROR CODE TO BLANK DISPLAY

:IS DATA A DECIMAL DIGIT?
:NO. KEEP ERROR CODE
;YES. MAKE DATA INTO A 16-81T INDEX
:GET BASE ADDRESS OF 7-SEGMENT TABLE
:FIND ELEMENT BY INDEXING
:GET 7-SEGMENT CODE FROM TABLE
:SAVE 7-SEGMENT CODE OR ERROR CODE

3FH
06H
5BH
4FH
66H
6DH
7DH
07H
7FH
6FH

library Example 3: Decimal Sum

Purpose: The program DECSUM adds two multi-word decimal numbers.
Language: Z80 assembler.
Initial Conditions: Address of LSBs of one number In Register Pair HL. address of LSBs
of other number in Register Pair DE. length of numbers (in bytes) In
A. Numbers arranged starting with LSBs at lowest address.
Final Conditions: Sum replaces number with starting address

In

Register Pair HL.

Requirements:
Memory
Time

11 bytes.
13 + 50N clock cvcles. where N is the number of
bytes involved.
Registers - A. B. D. E. H. L.
All flags affected. Carry shows if sum produced a carry.

Typical Case: (data

Start:
HL
DE
A
(0060l
(0061)
(0050)
(0051)
End:
(0060)
(0061)
CARRY

In

hexadecimall

0060
0050

2
34
55
88
15

22
71

15-12

Error Handling: Program does not check the validity of decimal Inputs. Accumulator
must be 1 or greater.
Listing:
DECSUM:
DECADD:

LD
AND
LD
ADC
DAA
LD
INC
INC
DJNZ
RET

B.A

A
A.(DE)
A.(HL)
(HL).A
DE
HL
DECADD

:COUNT = LENGTH OF NUMBERS (IN BYTES)

:CLEAR CARRY TO START
:GET 2 DECIMAL DIGITS FROM STRING 2
:ADD PAIR OF DIGITS FROM STRING 1
:MAKE ADDITION DECIMAL
:STORE RESULT IN STRING 1

TOTAL DOCUMENTATION
Complete documentation of microprocessor software will Include all or most of the elements that we have mentioned. So.
the total documentation package may involve:

DOCUMENTATION
PACKAGE

General flowcharts
A written description of the program
A list of all parameters and definitions
A memory map
A documented listing of the program
A description of the test plan and test results
The documentation may also include:
Programmers' flowcharts
Data flowcharts
Structured programs
The documentation procedures outlined above are the minimal acceptable set of
documents for non-production software. Production software demands even
greater documentation efforts. The following documents should also be produced:
Program Logic Manual
User GUide
Maintenance Manual
The program logic manual expands on the written explanation produced with the
software. It should be written for a technically competent individual who may not
possess the detailed knowledge assumed In the written explanation In the software.
The program logic manual should explain what the design goals of the system were.
what algOrithms were chosen to implement these goals. and what tradeoffs had to be
made in achieVing them.
It should then explain In great detail what data structures were employed and how they
are manipulated. It should provide a step-by-step gUide to the inner workings of the
code. Finally. it should contain any special tables or graphs that help explain any of the
concepts embodied in the code. Code conversion charts. state diagrams. translation
matrices. and flowcharts should be Included.
The user guide is probably the most important and most overlooked piece of documentation. No matter how well a system is designed. it is useless If no one can
use it effectively. The user guide should provide all users. sophisticated and unsophisticated. with an introduction to the system. It should then proYide detailed ex-

15-13

planations of system features and their use. Use plenty of examples because a good example can crystallize the information contained In many pages of text. Step-by-step
directions should be given. Test the user guide. I.e.. tryout the step-by-step usage procedures as yOU have documented them. Programmers with detailed knowledge of a
system's design often take shortcuts that are not at all apparent to the general reader.
An entire book could be written about the writing of user guides. and further diSCUSSion
IS beyond the scope of this book. However. remember that yOU can never spend too
much effort In preparing a user gUide. because it will be the most used of all system
documents.

The maintenance manual is designed for the programmer who has to modify the
system. It should outline step-by-step procedures for those reconfiguratlons designed
into the system. In addition. It shou Id outline any prOVISions placed Into code for future
expansion.
Documentation should not be taken lightly or postponed until the end of the software development. Proper documentation. combined with proper programming
practices. is not only an important part of the final product but can also make
development simpler. faster. and more productive. The designer should make consistent and thorough documentation part of every stage of software development.

REDESIGN
Sometimes the designer may have to squeeze the last microsecond of speed or
the last byte of extra memory out of a program. As larger Single-chip memories have
become available. the memory problem has become less serious. The time problem. of
course. is serious only if the application IS time-critical: in many applications the
microprocessor spends most of its time waiting for external devices. and program speed
IS not a malor factor.
Squeezing the last bit of performance out of a program is
seldom as important as some writers would have you believe.
In the first place. the practice is expensive for the follOWing
reasons:
1)
2)
3)
4)

It requires extra programmer time. which is often the Single largest cost in software
development.
It sacrifices structure and Simplicity with a resulting Increase In debugging and
testing time.
The programs require extra documentation.
The resulting programs will be difficult to extend. maintain. or re-use.

In the second place. the lower per-unit cost and higher performance may not really
be important. Will the lower cost and higher performance really sell more units? Or
would yOU do better with more user-oriented features? The only applications that
would seem to justify the extra effort and time are very high-volume. low-cost
and low-performance applications where the cost of an extra memory chip will far
outweigh the cost of the extra software development. For other applications. you
will find that you are playing an expensive game for no reason.

------.....

However. if you must redesign a program. the following

MAJOR OR
hints will help. First. determine how much more perforMINOR
mance or how much less memory usage is necessary. If
REORGANIZATION
the required improvement is 25% or less. you may be
able to achieve it by reorganizing the program. If it is more than 25%. you have
made a basic design error; you will need to consider drastic changes in hardware
or software. We will deal first with reorganization and later With drastic changes. You
should also look at Chapter 5 of
Programming for Logic Design for some examples.

zao

15-14

Note particularly that saving memory can be critical if it allows a program to fit Into the
limited amount of ROM and RAM available in a Simple one-chip or two-chip microcomputer. The hardware cost for small systems can thus be substantially reduced. if their
requirements can be limited to the memory size and I/O limitations of that particular
one-chip or two-chip system.

REORGANIZING TO USE lESS MEMORY

The following procedures will reduce memory usage for Z80
assembly language programs:
1)

Replace repetitious in-line code with subroutines. Be

sure. however. that the CALL and RETURN instructions do not offset most of the
gain. Note that this replacement usually results in slower programs because of the
time spent in transferring control back and forth.

2)

Use register operations when possible. But remember the cost of the extra inItialization.

3)

Use the Stack when possible. The Stack Pointer IS automatically updated after
each use so that no explicit updating instructions are necessary.

4)

Eliminate Jump instructions. Try to reorganize the program or use indirect Jumps
(JP (HU or JP (IX or IY)). RST. or RETURN instructions.

5)

Take advantage of addresses that you can manipulate as 8-bit quantities.

These include page zero and addresses that are multiples of 100 hexadeCimal. For
example. you might try to place all ROM tables in one lO016-byte section of
memory. and all RAM variables into another lO016-byte section.

6)

Organize data and tables so that you can address them without worrying
about address calculation carries or without any actual indexing. ThiS will
again allow you to manipulate 16-blt addresses as 8-bit quantities. See pages 5-1
to 5-6 of Z80 Programming for Logic Design for an example.

7)

Use the 16-bit instructions to replace two separate 8-bit operations. ThiS
may be particularly useful in initialization or storing results.

8)

Use leftover results from previous sections of the program.

9)

Take advantage of such instructions as INC (HU. OCR (HU. LD (HU. RL (HU. and
RR (HU. which operate directly on memory locations without using registers.

10)

Use INC or DEC to set or reset flag bits.

11)

Use relative jumps rather than jumps with direct addressing.

12)

Take advantage of the Block Move. Block Search. and Block I/O instructions
whenever you are handling blocks of data.

13)

Watch for special short forms of instructions such as the Accumulator shifts
(RLCA. RLA. RRCA. and RRA) and DJNZ.

14)

Use algorithms rather than tables to calculate arithmetic or logical expressions

and to perform code conversions. Note that this replacement may result in slower
programs.

15)

Reduce the size of mathematical tables by interpolating between entries. Here

again. we are saving memory at the cost of execution time.

16)

Take advantage of the alternate register set to cut down on the use of
storage. This can save time as well.

Although some of the methods that reduce memory usage also

save time. you can generally save an appreciable amount of
time only by concentrating on frequently executed loops. Even

15-15

SAVING
EXECUTION
TIME

completely eliminating an instruction that is executed only once can save at most a few
microseconds. But a savings in a loop that is executed frequently will be multiplied
many times over.
So, if you must reduce execution time. proceed as follows:
1)

Determine how frequently each program loop is executed. You can do this by
hand or by uSing the software simulator or another testing method.

2)

Examine the loops in the order determined by their frequency of execution.

starting with the most frequent. Continue through the list until you achieve the required reduction.

3)

First. see if there are any operations that can be moved outside the loop. I.e..
repetitive calculations, data that can be placed into a register or the Stack, addresses that can be placed into register pairs or Index registers, special cases or
errors that can be handled elsewhere, etc. Note that this will require extra inItialization and memory but will save time.

4)

Try to eliminate Jump statements. These are very time-consuming. Or. use
Jumps with direct addressing that require more memory but less time than jumps
with relative addressing.

5)

Replace subroutines with in-line code. This will save at least a CALL and a
RETURN Instruction.

6)

Use the Stack for temporary data storage.

7)

Use any of the hints mentioned in saving memory that also decrease execution time. These Include the use of block handling instructions, B-bit addresses,
16-blt instructions, RST, special short forms of instructions, etc.

B)

Do not even look at instructions that are executed only once. Any changes
that you make In such instructions only inVite errors for no appreciable gain.

9)

Avoid indexed and relative addressing whenever possible because they take
extra time.

10)

Use tables rather than algorithms; make the tables handle as much of the tasks
as possible even if many entries must be repeated.

MAJOR REORGANIZATIONS
If you need more than a 26% increase in speed or decrease in memory usage. do
not try reorganizing the code. Your chances of getting that much of an improvement are small unless you call in an outside expert. You are generally better off
making a major change.
The most obvious change is a better algorithm. Particularly if
BETTER
you are doing sorts, searches, or mathematical calculations, you
ALGORITHMS
may be able to find a faster or shorter method in the literature.
libraries of algorithms are available in some lournals and from professional groups. See,
for example, References 1 through 10 at the end of this chapter.
More hardware can replace some of the software. Counters. shift registers,
arithmetic units, hardware multipliers, and other fast add-ons can save both time and
memory. Calculators, UARTs, keyboards, encoders, and other slower add-ons may save
memory even though they operate slowly. Compatible parallel and serial interfaces, and
other devices specially designed for use with the ZBO may save time by taking some of
the bu rden off the CPU.

15-16

Other changes may help as well:

OTHER
MAJOR
CHANGES

1)

A CPU with a longer word will be faster if the data is long

enough. Such a CPU will use less total memory. 16-bit processors. for example. use memory more effiCiently than 8-bit
processors. since more of their Instructions are one word long.

2)

Versions of the CPU may exist that operate at higher clock rates. But remember that you will need faster memory and I/O ports. and you will have to adjust any
delay loops.

3)

Two CPUs may be able to do the job in parallel or separately if you can diVide the
Job and solve the communications problem.

4)

A specially microprogrammed processor may be able to execute the same program much faster. The cost. however. will be much higher even if you use an offthe-shelf emu lation.

5)

You can make tradeoffs between time and memory. Lookup tables and function
ROMs will be faster than algorithms. but will OCCUPy more memory.

F-"---...,.

This kind of problem. in which a large improvement is necesDECIDING

sary. usually results from lack of adequate planning in the
ON A MAJOR
definition and design stages. In the problem definition stage
CHANGE
you should determine which processor and methods will be
adequate to handle the problem. If you misjudge. the cost later will be high. A
cheap solution may result in an unwarranted expenditure of expensive development time. Do not try to just get by; the best solution is usually to do the proper
design and chalk a failure up to experience. If you have followed such methods as
flowcharting. modular programming. structured programming. top-down design.
and proper documentation. you will be able to salvage a lot of your effort even if
you have to make a major change.

15-17

REFERENCES
1.

Collected AlgOrithms from ACM. ACM, Inc.. P O. Box 12105, Church Street StatIOn, New York 10249.

2.

Chen, T. C.. "Automatic Computation of Exponentials, Logarithms, Ratios, and

Square Roots," IBM Journal of Research and DeveIOJ:.1ment. Volume 18. pp.
380-388, July, 1972.
--

3.

H. Schmid, Decimal ComRutation, Wiley-Intersclence, New York, 1974.

4.

Knuth, D. E.. The Art of ComRuter Programming, Volume 1. Fundamental

.tiJgorithms,Addison-Wesley, Reading, Mass.. 1967

5.

Knuth, D. E.. The Art of ComRuter Programming, Volume 2: Semlnumerical

Alg~, Addison-Wesley, Reading, Mass.. 1969.

6.

Knuth, D. E.. The Art of Computer Programming, Volume 3: Sorting and Search-

llli!' Addison-Wesley, Reading, Mass.. 1973.

al..~P..Rlied

Carnahan, B. et

S.

Despain, A. M. "FOUrier Transform Computers USing CORDIC Iterations," IEEE

Transactions on ComRuters, October 1974. pp. 993-1001.

9.

Luke, Y L.. AlgOrithms for the ComRutation of Mathematical Functions, AcademiC

Press, New York, 1977

Numerical Methods, Wiley, New York, 1969.

10.

Hwang, K.. ComRuter Arithmetic, Wiley, New York, 1978.

11.

Dollhoff, T.. "Microprocessor Software: How to Optimize Timing and Memory

Usage. Part Four: Techniques for the Zilog Z80," Digital DeSign, February 1977.
pp.44-51.
- --- -

15-18

Chapter 16
SAMPLE PROJECTS
PROJECT #1: A Digital Stopwatch
Purpose: This project is a digital stopwatch. The operator enters
STOPWATCH
INPUT
two digits (minutes and tenths of minutes) from a
PROCEDURE
calculator-like kevboard and then presses the GO kev.
The system counts down the remaining time on two
seven-segment LED displavs (see Chapter 11 for a description of unencoded
keyboards and LED displays).
Hardware: The project uses one input port and one output port (one Z80 Parallel
Input/Output Device or PIOl. two seven-segment displays. a 12-kev kevboard. a 7404
Inverter. and either a 7400 NAND gate or a 7408 AND gate. depending on the polantv
of the seven-segment displays. The displays may require drivers. inverters. and resIstors. depending on their polantv and configuration.

The hardware is organized as shown In Figure 16-1. Output lines O. 1, and 2 are used to
scan the kevboard. Input lines O. 1. 2, and 3 are used to determine whether anv keys
have been pressed. Output lines 0, 1, 2, and 3 are used to send BCD digits to the sevensegment decod.er/dnvers. Output line 4 IS used to activate the LED displays (if line 4 IS
'1', the displays are lit). Output line 5 is used to select the left or right displav: output
line 5 is '1' if the left display IS being used, '0' if the right display is being used. Thus.
the common line on the left display should be active if line 4 is '1' and line 5 is '1', while
the common line on the right display should be active if line 4 is '1' and line 5 IS '0'
Output line 6 controls the right-hand decimal POint on the left display. It may be dnven
with an inverter or simplv left on.
Keyboard Connections: The keyboard IS a Simple calculator keyboard available for
50 from a local source. It consists of 12 unencoded key-switches arranged in four rows
of three columns each. Since the wiring of the keyboard does not cOincide with the observed rows and columns. the program uses a table to Identifv the keys. Tables 16-1
and 16-2 contain the Input and output connections for the keyboard. The deCimal point
key is present for operator convenience and for future expansion: the current program
does not actually use the kev.

In an actual application. the kevboard would require pullup resistors to ensure that the
inputs would actuallv be read as logic '1's when the keys were not being pressed. It
would also require current-limiting resistors or diodes on the output port to avoid
damaging the drivers In the case where two outputs were dnvlng against each other.
This could occur if two kevs in the same row were pressed at the same time. thus connecting two different column outputs.

16-1

87 r--(not usedl
8e
Output 8S
Port
84
(PIO
83
Port 81 82
81

So

II

Co C1 c 2

Input
Port
(Pia
Port AI

A3

R3

A2

R2

A1

R1

Ao

RO

KeYboard

00 0 1 02 0 3

OP
Display
and

If

DO 0 1 02 0 3
Display
and

Driver

Driyer

Hefti

(rightl

Common

Common

lL=D

lb

Figure 16-1, Digital Stopwatch I/O Configuration

Table 16-1, Input Connections for Stopwatch Keyboard

Kevs Connected

Input Bit

'3', '5', 'S'

'2' '6', 'g'
'0', '1', '7'
'4' '.. , 'GO'

1
2

Table 16-2, Output Connections for Stopwatch KeYboard

Output Bit

Keys Connected
'0', '2', '3', '4'
'1' 'S', 'g' 'GO'
'5', '6', '1', '..

0
1
2

16-2

General Program Flowchart:

Initialization

Identify
key closure

Save key value

Count time on LEOs

16-3

Display Connections: The displays are seven-segment displays with their own integral decoders. A typical example would be the Texas Instruments TIL309 device.
which has an Internal TTL MSI chip with latch. decoder. and driver. Clearly. standard
seven-segment displays would be cheaper but would require some additional software
(the seven-segment conversion routine shown In Chapter 7). Data is entered into the
display as a single binary coded deCimal digit; the digits are represented as shown in
Figure 11-15. The deCimal POint is a single LED that is turned on when the deCimal
POint Input is a logic '1' You can find more information about displays in References 10
and 11 at the end of this chapter.
Program Description:
The program IS modular and has several subroutines. The emphaSIS IS on clanty and
generality rather than efficiency; obViously. the program does not utilize the full
capabilities of the Z80 processor. Each section of the listing will now be described In
detail.
1)

Introductory Comments
The Introductory comments fully describe the program; these comments are a
reference so that other users can easily apply. extend. and understand the program. Standard formats. indentations. and spacings Increase the readability of the
program.

2)

Variable Definitions
All vanable definitions are placed at the start of the program so that they can easily
be checked and changed. Each vanable IS placed in a list alphabetically With other
variables of the same type; comments describe the meaning of each vanable. The
categones are;
a)

Memory system constants that may vary from system to system depending on
the memory space allocated to different programs or types of memones

b)

Temporary storage (RAM) used for vanables

cl

I/O (PIO) port addresses

Definitions

d)

The memory system constants are placed In the definitions so that the user may
relocate the program. temporary storage. and memory stack without making any
other changes. The memory constants can be changed to accommodate other
programs or to cOincide With a particular system's allocation of ROM and RAM addresses.
Temporary storage IS allocated by means of DEFS (Define Storage) pseudo-operations. An ORG (ongln) pseudo-operation places the temporary storage locations in
a particular part of memory. No values are placed in these locations so that the
program could eventually be placed In ROM or PROM and the system could be
operated from power-on reset Without reloading.
Each port address occupied by a PIO is named so that the addresses can easily be
changed to handle vaned configurations. The naming also serves to clearly distinguish control registers from data registers.
The definitions clarify the meaning of certain constants and allow parameters to
be changed easily. Each definition is given in the form (binary. hex. octal. ASCII. or
decimailin which ItS meaning IS the clearest. Parameters (such as debounce time)
are placed here so that they can be varied With system needs.

16-4

3)

Initialization

zao

Memory location 0 (the reset location on the

microprocessor) contains a lump
to the starting address of the main program. The main program can thus be
placed anywhere in memory and reached via a "RESET" signal.
The initialization consists of four steps:
al

4)

Place a starting value In the Stack Pointer. The Stack IS used only to store
subroutine return addresses.

b)

Configure the PIO control registers.

cl

Start the number of digit keys pressed at zero.

dl

Initialize the location where the next digit key pressed will be saved to the
start of the digit key array. An indirect procedure IS used. in which KEYAD
contains the address in which the next digit will be placed. Each time a digit
key is recognized. the contents of KEYAD are incremented so that the next
digit key will be placed Into the next memory location.

Look for Key Closure

Flowchart:

Ground all keyboard

columns

Key closures are Identified by grounding all the keyboard columns and then
checking for grounded rows (j.e.. column-to-row sWitch closures). Note that the
program does not assume that the unused input bits are all high: Instead. the bits
attached to the keyboard are isolated with a logical AND instruction.
5)

Debounce Key
The program debounces the key closure in software by waiting for two milliseconds. This is usually long enough for a clean contact to be made. Subroutine
DELAY Simply counts with Register C for 1 millisecond. The number of milliseconds IS In the Accumulator. DELAY would have to be adjusted if a slower clock or
slower memories were being used. You could make the change simply by redefinIng the constant MSCNT.

16-5

6)

Identify Key Closure

Flowchart:

Ground a keyboard
column bV output of
(pattern pOinter)

Yes

"crement BV ta -e

Increment key teble

pointer by 1
Shift keyboar,f input
right 1 bit

pointer by the numba

of keys in a column
(KeOL) Increment
attern pointer b 1

Key 10 =
(key table pointer)
Use key table pOinter
to get key 10

The particular key closed IS Identified by grounding single columns and observing
whether a closure is found. Once a closure IS found (so the key column is known).
the key row can be determined by shifting the input.
The patterns required to ground single keyboard columns are in a table PATT In
memory. The final pattern In the table IS a marker (ECODE) which Indicates that all
the columns have been grounded Without a closure being found. This pattern also
Indicates to the main program that the closure could not be identified (e.g.. the
key closure ended or a hardware error occurred before we could find the closure).

16-6

IKEY1

The key Identifications are in table KTAB In memory. The

keys In the first column (attached to the least significant out~
put bit) are followed by those In the second column. etc.
Within a column. the key In the row attached to the least Significant Input bit is
first. etc. Thus. each time a column is scanned without finding a closure. the number of keys in a column (NROWS) must be added to the key table pOinter In order
to move to the next column. The key table pointer is also Incremented by one
before each bit in the row Inputs IS examined: this process stops when a zero input
IS found. Note that the key table pointer IS started one location before the table.
since it is always Incremented once in the search for the proper row.
If we cannot Identify the key closure. we simplY Ignore It and look for another
closure.
7)

Act on Key Identification

If the program has enough digits (two In thiS simple caseL It looks only for the GO
key and Ignores all other keys. If it finds a digit key. it saves the value In the key
array. increments the number of digit keys pressed. and Increments the key array
pOinter.
If the entry is not complete. the program must wait for the key closure to end so
that the system will not read the same closure again. The user must wait between
key closures (i.e.. release one key before pressing another one!' Note that the program will Identify double key closures as one key or the other. depending on
which closure the Identification routine finds first. An improved version of thiS
program would display digits as they were entered and would allow the user to
omit a leading or trailing zero. (i.e.. key in ..... "7". "GO" to get a count of seventenths of a minute!'

8)

Set Up Display Output

The digits are placed in registers or memorv locations With bit 4 set so that the
output is' sent to the displays. Bits 5 and 6 are set for the most significant digit to
direct the output to the left display and to turn on the deCimal point.

9)

Pulse the LED Displays

Each display is turned on for two milliseconds. ThiS process IS repeated 1500
times in order to get a total delay of 0.1 minutes. or 6 seconds. The pulses are frequent enough so that the LED displays appear to be lit continuously.

16-7

1O}

Decrement Display Count

Flowchart:

Right Display

= Right

Display - 1

Left Display =

Left Display - 1

Yes

Right Display

End of timer

=9

program

The value of the less significant digit is reduced by one. If this affects bit 4
(LEDON - used to turn the displays on), the digit has become negative. A borrow
must then be obtained from the more significant digit. If the borrow from the more
significant digit affects bit 4. the count has gone past zero and the countdown is
finished. Otherwise, the program sets the value of the less Significant digit to 9
and continues.
Note that comments describe both sections of the program and indiVidual statements.
The comments explain what the program IS dOing, not what specific Instruction codes
do. Spacing and indentation have been used to Improve readability.

16-8

;PROGRAM NAME: TIMER

;DATE OF PROGRAM: 10/24/78
;PROGRAMMER; LANCE A. LEVENTHAL
;PROGRAM REQUIREMENTS: D1 (209) BYTES
;RAM REQUIREMENTS: 5 BYTES
;1/0 REQUIREMENTS: 1 INPUT PORT. 1 OUTPUT PORT (1

zao PIO)

.THIS PROGRAM IS A SOFTWARE TIMER WHICH ACCEPTS INPUTS FROM A

CALCULATOR-LIKE KEYBOARD AND THEN PROVIDES A STOPWATCH
COUNTDOWN ON TWO 7-SEGMENT LED DISPLAYS IN MINUTES AND TENTHS
OF MINUTES
;KEYBOARD
;A 12-KEY KEYBOARD IS ASSUMED
.THREE COLUMN CONNECTIONS ARE OUTPUTS FROM THE PROCESSOR
SO THAT A COLUMN OF KEYS CAN BE GROUNDED
;FOUR ROW CONNECTIONS ARE INPUTS TO THE PROCESSOR SO THAT
COMPLETED CIRCUITS CAN BE IDENTIFIED
;THE KEYBOARD IS DEBOUNCED BY WAITING FOR TWO MILLISECONDS
AFTER A KEY CLOSURE IS RECOGNIZED
;A NEW KEY CLOSURE IS IDENTIFIED BY WAITING FOR THE OLD ONE
TO END SINCE NO STROBE IS USED
:THE KEYBOARD COLUMNS ARE CONNECTED TO BITS 0
TO 2 OF THE PIO B PORT
THE KEYBOARD ROWS ARE CONNECTED TO BITS 0
TO 3 OF THE PIO A PORT
;DISPLAYS
;TWO 7-SEGMENT LED DISPLAYS ARE USED WITH SEPARATE DECODERS
(7447 OR 7448 DEPENDING ON THE TYPE OF DISPLAY)
.THE DECODER DATA INPUTS ARE CONNECTED TO BITS 0 TO 3
OF THE PIO B PORT
;BIT 4 OF THE PIO B PORT IS USED TO ACTIVATE THE LED
DISPLAYS (BIT 4 IS 1 TO SEND DATA TO LEDS)
;BIT 5 OF THE PIO B PORT IS USED TO SELECT WHICH
LED IS BEING USED (BIT 5 IS 1 IF THE LEADING DISPLAY
IS BEING USED. 0 IF THE TRAILING DISPLAY IS BEING USED)
;BIT 6 OF THE PIO B PORT IS USED TO LIGHT THE DECIMAL
POINT LED ON THE LEADING DISPLAY (BIT 6 IS 1 IF
THE DISPLAY IS TO BE LIT)
;METHOD
;STEP 1 - INITIALIZATION
THE MEMORY STACK POINTER (USED FOR SUBROUTINE RETURN
ADDRESSES) IS INITIALIZED. THE NUMBER OF DIGIT KEYS PRESSED IS SET
TO ZERO. AND THE ADDRESS INTO WHICH THE NEXT DIGIT KEY
IDENTIFICATION WILL BE PLACED IS INITIALIZED TO THE FIRST ADDRESS
IN THE DIGIT KEY ARRAY
;STEP 2 - LOOK FOR KEY CLOSURE
ALL KEYBOARD COLUMNS ARE GROUNDED AND THE KEYBOARD ROWS
ARE EXAMINED UNTIL A CLOSED CIRCUIT IS FOUND

16-9

:STEP 3 - DEBOUNCE KEY CLOSURE

A WAIT OF 2 MS IS INTRODUCED TO ELIMINATE KEY BOUNCE
:STEP 4 - IDENTIFY KEY CLOSURE
THE KEY CLOSURE IS IDENTIFIED BY GROUNDING SINGLE KEYBOARD
COLUMNS AND DETERMINING THE ROW AND COLUMN OF THE KEY
CLOSURE. A TABLE IS USED TO ENCODE THE KEYS ACCORDING TO THEIR
ROW AND COLUMN NUMBER
IN THE KEY TABLE, THE DIGITS ARE IDENTIFIED BY THEIR VALUES,
THE DECIMAL POINT KEY IS NO, 10, AND THE "GO" KEY IS NO, 11
:STEP 5 - SAVE KEY CLOSURE
DIGIT KEY CLOSURES ARE SAVED IN THE DIGIT KEY ARRAY UNTIL
TWO DIGITS HAVE BEEN IDENTIFIED, DECIMAL POINTS, FURTHER DIGITS,
AND CLOSURES OF THE "GO" KEY BEFORE TWO DIGITS HAVE BEEN
IDENTIFIED ARE IGNORED
AFTER TWO DIGITS HAVE BEEN FOUND, THE "GO" KEY IS USED TO
START THE COUNTDOWN PROCESS
:STEP 6 - COUNT DOWN TIMER INTERVAL ON LEDS
A COUNTDOWN IS PERFORMED ON THE LEOS WITH THE LEADING DIGIT
REPRESENTING THE REMAINING NUMBER OF MINUTES AND THE TRAILING
DIGIT REPRESENTING THE REMAINING NUMBER OF TENTHS OF MINUTES

TIMER VARIABLE DEFINITIONS

:MEMORY SYSTEM CONSTANTS
BEGIN

EQU

EQU
LASTM
TEMP
EQU
:RAM TEMPORARY
ORG
KEYAD: DEFS

KEYNO:

DEFS

NKEYS:

DEFS

50H
1000H
800H
STORAGE
TEMP
2

:BEGIN IS STARTING MEMORY LOCATION

FOR PROG
:LASTM IS STARTING STACK ADDRESS
:TEMP IS START OF RAM STORAGE

:KEYAD HOLDS THE ADDRESS IN THE

DIGIT KEY ARRAY IN WHICH THE
IDENTIFICATION OF THE NEXT DIGIT
, KEY WILL BE PLACED
:KEYNO IS THE DIGIT KEY ARRAY - IT
, HOLDS THE IDENTIFICATIONS OF THE
, DIGIT KEYS THAT HAVE BEEN PRESSED
:NKEYS HOLDS NUMBER OF DIGIT KEYS
, PRESSED

:1/0 UNITS AND PIO ADDRESSES

PIODRA
PIOCRA
PIODRB

EQU
EQU
EQU

OEOH
OE2H
OE1H

PIOCRB

EQU

OE3H

:INPUT PIO FOR KEYBOARD

:OUTPUT PIO FOR KEYBOARD AND
, DISPLAY

:DEFINITIONS
DECPT

EQU

;BIT POSITION TO TURN ON DECIMAL

POINT LED

16-10

ECODE

EQU

OFFH

GOKEY
LEDON
LEDSL

EQU
EQU
EQU

11
4

MSCNT
MXKEY

EQU
EQU

OF9H
2

NROWS

EQU

OPEN

EQU

00001111B

TPULS
TWAIT

EQU
EQU

2
2

ORG

;ERROR CODE IF 10 ROUTINE DOES NOT FIND

KEY
;IDENTIFICATION NUMBER FOR "GO" KEY
;BIT POSITION TO SEND OUTPUT TO LEOS
;BIT POSITION TO SELECT LEADING
DISPLAY
:COUNT NEEDED TO GIVE 1 MS DELAY TIME
:MAXIMUM NUMBER OF DIGIT KEY
CLOSURES USED
;NUMBER OF ROWS IN KEYBOARD OR KEYS
IN COLUMN
;INPUT FROM KEYBOARD IF NO KEY
CLOSED
:NUMBER OF MS BETWEEN DIGIT DISPLAYS
:NUMBER OF MS TO DEBOUNCE KEYS

;RESET ROUTINE TO REACH TIMER PROGRAM

JP

BEGIN

;FIND TIMER PROGRAM

;INITIALIZATION OF TIMER PROGRAM

ORG
LD
OUT
LD
OUT
LD
SUB
LD
LD
LD

BEGIN
A.01001111B
(PIOCRA).A
A.00001111 B
(PIOCRBLA
SP.LASTM
A
(NKEYS).A
HL.KEYNO
(KEYAD).HL

;MAKE PIO PORT A INPUT

;MAKE PIO PORT B OUTPUT
;PUT STACK AT END OF MEMORY
:NUMBER OF DIGIT KEYS PRESSED = ZERO
;STARTING LOCATION FOR DIGIT KEYS

;SCAN KEYBOARD LOOKING FOR KEY CLOSURE

START

CALL

SCANC

;WAIT FOR KEY CLOSURE

;WAIT FOR KEY TO BE DEBOUNCED

LD
CALL

A.TWAIT
DELAY

:GET DEBOUNCE TIME IN MS

:WAIT FOR KEY TO STOP BOUNCING

;IDENTIFY WHICH KEY WAS PRESSED

CALL
CP
JR

IDKEY
ECODE
Z.START

IDENTIFY KEY CLOSURE

WAS KEY CLOSURE IDENTIFIED?
NO. WAIT FOR ANOTHER CLOSURE

;ACT ON KEY IDENTIFICATION

16-11

LD
LD

B,A
HL,NKEYS

LD
CP
JR
LD
CP
JR
INC
LD
LD
INC
LQ

A,(HU
MXKEY
Z,KEYF
A,B

10
NC,WAITK
(HU
HL,(KEYAD)
(HU,A
HL
(KEYADl.HL

;SAVE KEY NUMBER

;CHECK FOR MAXIMUM NUMBER OF DIGIT
KEYS
;HAS MAXIMUM BEEN REACHED?
;YES, LOOK FOR GO KEY
;NO, LOOK FOR DIGIT KEYS ONLY
;IS THIS KEY A DIGITI
;NO, IGNORE IT
;YES, INCREMENT DIGIT KEY COUNTER
;SAVE KEY NUMBER IN ARRAY

:WAIT FOR CURRENT KEY CLOSURE TO END

WAITK:

CALL
JR

SCANO
START

;WAIT FOR KEY TO BE RELEASED

;GO LOOK FOR NEXT KEY

;LOOK FOR GO KEY IF ENOUGH DIGITS FOUND

KEYF

LD
CP
JR

A,B
GOKEY
NZ,WAITK

;GET NUMBER OF KEY PRESSED

;IS IT "GO" KEY?
;NO, IGNORE IT

;PUT DIGITS INTO REGISTERS FOR DISPLAY

LD
LD
SET
SET
SET
INC
LD
SET

HL,KEYNO
D,(HU
DECPT,D
LEDON,D
LEDSL,D
HL
E,(HL)

LEDON,E

;GET LEADING DIGIT

;TURN ON DECIMAL POINT
;SET OUTPUT TO LEOS
;SELECT LEADING DISPLAY
;GET TRAILING DIGIT
;SET OUTPUT TO LEOS

;PULSE THE LED DISPLAYS

LEDLP'
TLOOP:
LDPUL.

LD
LD
LD
OUT
LD
CALL
OUT
LD
CALL
DJNZ
DEC
JR

C,PIODRB
H,6
B,250

;GET OUTPUT PORT ADDRESS

;SET COUNTERS FOR 6 SECONDS

(C),D

;OUTPUT LEADING DIGIT TO LED 1

;DELAY BETWEEN DIGITS

A.TPULS
DELAY
(Cl.E
A.TPULS
DELAY
LDPUL

;OUTPUT TRAILING DIGIT TO LED

;DELAY BETWEEN DIGITS

H
NZ.TLOOP

;DECREMENT COUNT ON LED DISPLAYS

16-12

DEC
BIT

JR
DEC
BIT
JP
LD
SET

JR

E
LEDON.E
NZ.LEDLP
D

LEDON.D
Z.BEGIN
E.9
LEDON.E
LEDLP

COUNT DOWN TRAILING DIGIT

IS TRAILING DIGIT PAST ZERO?
NO. CONTINUE
COUNT DOWN LEADING DIGIT
:IS LEADING DIGIT PAST ZERO?
;YES. WAIT FOR NEXT TIMING TASK
;NO. SET TRAILING DIGIT TO 9
:SET OUTPUT TO LEDS
:RETURN TO DISPLAY SECTION

;SUBROUTINE SCANC SCANS THE KEYBOARD WAITING FOR A KEY CLOSURE

;ALL KEYBOARD INPUTS ARE GROUNDED
SCANC;

SUB
OUT
IN
AND
CP

JR

A
(PIODRBl.A
A.(PIODRA)
OPEN
OPEN
Z.SCANC

;GROUND ALL KEYBOARD COLUMNS

;IGNORE UNUSED INPUTS

;ARE ANY KEYS CLOSED?
;NO. CONTINUE SCANNING

RET
;SUBROUTINE DELAY WAITS FOR THE NUMBER OF MILLISECONDS SPECIFIED
. IN REGISTER A
DELAY
DLYl
WTLP'

EXX
LD
DEC

JR
DEC

JR

C.MSCNT
C
NZ.WTLP
A
NZ.DLYl

EXX
RET

;SAVE USER REGISTERS

;LOAD REGISTER C FOR 1 MS
:WAIT 1 MS
;COUNT DOWN NUMBER OF MS
;RESTORE USER REGISTERS

;SUBROUTINE IDKEY DETERMINES THE ROW AND COLUMN NUMBER OF THE

. KEY CLOSURE AND IDENTIFIES THE KEY BY USING A TABLE
IDKEY

LD
LD
LD

BC.PATT
HL.KTAB-l
DE.NROWS

;POINT TO SCAN PATTERNS

;START KEY TABLE POINTER
;GET NUMBER OF KEYS IN A COLUMN

;SCAN KEYBOARD COLUMNS SUCCESSIVELY LOOKING FOR CLOSURE

FCOL.

LD
CP
RET
OUT
IN
AND
CP

JR
ADD
INC

JR

A.(BC)
ECODE
Z
(PIODRBl.A
A,(PIODRA)
OPEN
OPEN
NZ.FROW
HL.DE
BC
FCOL

:GET PATTERN TO GROUND COLUMN

;ALL COLUMNS SCANNED?
:YES. RETURN WITH ERROR CODE
;SCAN COLUMN
;IGNORE UNUSED INPUTS
;ANY KEYS IN THIS COLUMN CLOSED?
YES. GO DETERMINE CLOSURE ROW
NO. MOVE KEY TABLE POINTER TO
NEXT COLUMN
POINT TO NEXT SCAN PATTERN

16-13

:DETERMINE ROW NUMBER OF CLOSURE

FROW'

INC
RRCA

HL

JR

C,FROW

:MOVE KEY TABLE POINTER TO NEXT ROW

:NEXT ROW GROUNDED?
:NO, KEEP LOOKING

:IDENTIFY KEY FROM TABLE

LD
RET

A,(HL)

:GET KEY NUMBER

:SCAN PATTERNS USED TO GROUND ONE COLUMN AT A TIME

:ERROR PATTERN USED TO INDICATE THAT ALL COLUMNS HAVE BEEN SCANNED
:THE COLUMN ATTACHED TO OUTPUT BIT 0 IS SCANNED FIRST, THEN
THE ONE ATTACHED TO OUTPUT BIT 1. ETC.
PATT

DEFB
DEFB
DEFB
DEFB

00000110B
00000101B
00000011B
ECODE

:KEYBOARD TABLE
:COLUMNS ARE PRIMARY INDEX, ROWS SECONDARY INDEX
:THE KEYS IN THE COLUMN ATTACHED TO OUTPUT BIT 0 ARE FOLLOWED
BY THOSE IN THE COLUMN ATTACHED TO OUTPUT BIT 1, ETC. WITHIN
A COLUMN. THE KEY ATTACHED TO INPUT BIT 0 IS FIRST FOLLOWED
. BY THE ONE ATTACHED TO INPUT BIT 1. ETC.
:THE DIGIT KEYS ARE 0 TO 9, DECIMAL POINT IS 10, GO IS 11
KTAB:

DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB

3
2
0

4
8
9
1
11

5
6
7
10

:CO,RO
:CO,R1
:CO,R2
:CO,R3
:C1.RO
:C1.R1
:C1.R2
:C l.R3
:C2.RO
:C2.R1
:C2,R2
:C2,R3

:SUBROUTINE SCANO SCANS THE KEYBOARD WAITING FOR KEY CLOSURE TC

END SO NEXT CLOSURE CAN BE FOUND
SCANO:

SUB
OUT
IN
AND
CP

JR

:GROUND ALL KEYBOARD COLUMNS

(PIODRB),A
A,(PIODRA)
OPEN
OPEN
NZ,SCANO

;IGNORE UNUSED INPUTS

:ARE ANY KEYS STILL CLOSED?
:YES, CONTINUE SCANNING

RET
END

16-14

PROJECT #2: A Digital Thermometer

Purpose: This project IS a digital thermometer which shows the temperature in
degrees Celsius on two seven-segment displays.
Hardware: The project uses one input port and one output port. two seven~segment
displays, a 74LS04 Inverter. a 74LSOO NAND gate or a 74LS08 AND gate depending on
the polarity of the displays. an Analog Devices AD7570J 8-bit monolithic AID converter, an LM311 comparator. and various peripheral drivers, resistors. and capacitors
as required by the displays and the converter. (See Chapter 11 and Reference 1 at the
end of this chapter for discussions of AID converters,)
Figure 16-2 shows the organization of the hardware. Output line 7 from PIO Port B IS
used to send a Start Conversion signal to the AID converter. Input lines 0 through 7 are
attached directly to the eight digital data lines from the converter. Output lines 0
through 3 are used to send BCD digits to the seven-segment decoder/drlvers. Output
line 4 activates the displays and output line 5 selects the left or right display (line 5 IS '1'
for the left display).
The analog part of the hardware is shown In Figure 16-3. The
THERMOMETER
ANALOG
thermistor simply provides a resistance that depends on temperature. Figure 16-4 IS a plot of the resistance and Figure 16-5
HARDWARE
shows the range of current values over which the resistance is
linear. The conversion to degrees Celsius in the program IS performed with a calibration
table. The two potentiometers can be adjusted to scale the data properly. A clock for
the AID converter is generated from an RC network. The values are R7=33 k!l and
C1 =1000 pF. so that the clock frequency IS about 75 kHz. At this frequency. the maximum conversion time for eight bits IS about 50 microseconds. A much longer delay is
allowed for conversion so that no check for the end of conversion is necessary. The 8bit version of the converter reqUires the follOWing special connections. The eight data
lines are DB2 through DB9 (DB1 IS always high during conversion and DBO low). The
Short Cycle 8-bit Input (pm 26-SC8) is tied low so that only an 8-blt conversion IS performed. In the present case, High Byte Enable (pin 20-HBEN) and Low Byte Enable (pin
21-LBENl were both tied high so that the data outputs were always enabled.
The AID converter uses the successive approximation method to perform a conversion.
The ADC's data register is connected to the mputs of an internal 01 A converter whose
output (available at OUT1 and OUT2) is compared to the analog input. When a conversion IS Initiated. the ADC logic sets the data register to all zeros with the exception of
the most significant bit (MSB). which is set to one. If the analog input IS less than the
resulting internally generated analog value. then the MSB IS reset to zero: otherWise It
remains a one. The next most significant bit IS then set to one and the process repeated
until all eight bits have been "tested" In thiS way. After the eighth cycle. the value m the
register IS the value which most closely corresponds to the analog input.
ThiS method IS fast. but it requires that the input be stable dUring the conversion process. Rapidly changing or nOIsy Inputs would require additional signal conditioning. The
references at the end of this chapter describe more accurate methods for handling
analog 1/0.

16-15

67

Output
Port
IPIO
Port 61

11-

"1

6S '--Inot usedl
6SI-

....

..,

64 1 - - - - - - t - - - l I . . - r - - - - - - - ,

63

1-----+--+-1-----......-1-+-----,

6Z II------I----t-+------Gj,...-t---t--t------,
6,

1------t----;-t--

BOI I - - - - - - - t - - - - , H - -

Ip-t-t---;-+---,

-j--t-t--t--j----,

Start
A7
Input

AS
AS

Port

A4

ConversIon

--f
--f
--f

(PIO A3 101111t---t
---t
Port AI AZ
A,
---t

Displav
and

Displav

AID
Converter

Dnver

Dover

lIeftl

(rlghtl

Ao -

Analog l n p u t - - - - - - - - - '

l~m""

and

Common

Figure 16-2. I/O Configuration for a Digital Thermometer

16-16

R6
50 kil
OFFSET ADJ

.."

.........

15Va

+5 V

+15V

VREF
10 V

R3
200 il

O+15V

R5
2Mil

22
2

R2
5kil

Voo

VCC

OUT1

VREF
+5 V

OUT2

24
CLOCK

COMP

+15 V
C1
JOOOPF
Rl

AD7570J
AID

SRO

Converter

SYNC

Thermistor

(not usedl
9
10

D89
3

ANALOG
INPUT

11

088

12

087
+5V

DB6
20
21

From PIO Port 8, bit 7

Inot usedl

13

25
28
27
26

HBEN

085

LBEN

084

14
15

(not usedl
A7
A6
A5
A4

To PIO
Port A

A3
A2

16

STRT

083

BUSY

DB2

BSEN

A1
17
18

DB1

19

5e8
AGND

AD
(not usedl
(not usedl

-::-0
Note: If positive VREF 's used, the ANALOG INPUT range '5 0 to ,VREF' and the
COMPARATOR's (.J Input should be connected to OUT1 Ipln 41 of the AD7570.
RT

IS

the thermistor. The analog mput from the vortage divider

IS:

R8
.15 Volt
R8 + RT
Since RF = 6B kil, the Input's:

1.02 Mil

Volt

RT + 68 kil
RT has a m'OImum value of 34 kil (T=50"C. see Figure 16-41 so full scale ,510 Volt.

Figure 16-3. Digital Thermometer Analog Hardware

16-17

l000000r-----------------------,

-- -- -100 000

T(OC)

R(Ohm)

0
25

365000
100 000
34 000
6000

50

-- ........

100

"Q.. .....

10000 ....

... ..............

..........

..........

-1-

....
.....

25

50

Temperature (OC)

Figure 16-4. Thermistor Characteristics

(Fenwal GA51Jl Beadl

100

r-------------------------,
The curve is linear (i.e.. the resistance is
independent of current) for currents less
than 0.1 milliampere.

10

0.01

0.1
I (milliampere)

Figure 16-5. TYPical E-I Curve for Thermistor (25Cl

16-18

1.0

General Program Flowchart'

Start

Initializaoon

-*
'I
Send Start
ConversIon S!gnal
to AID converter

t
Wait1ms

t
Read data from

AID converter

t
Convert data to
degrees Celsius

t
DispIav

temperature on

LEOs fa< six seconds

16-19

Program Description:
1)

Initialization
Location 0 (the Z80 microprocessor RESET location) contains a Jump to the starting
address of the main program.The initialization configures the PIO control registers
and starts the Stack POinter at the highest address in RAM. The Stack is used onlv
to store subroutine return addresses.

2)

Send START CONVERSION Signal to AID Converter

The CPU pulses the START CONVERSION line by first placing a '1' on line 7 of PIO
Port B and then placing a '0' on that line. Each input from the converter requires a
starting pu Ise.

3)

Wait 1 ms for Conversion

A delav of 1 ms after the START CONVERSION pulse guarantees a completed conversion. Actuallv, the converter takes onlv a maximum of 100 microseconds for an
8-bit conversion. We could reduce the delav by checking the BUSY signal from the
converter. This Signal is either a '1' (conversion complete) or '0' (conversion In
progress) if the BUSY ENABLE line is addressed. In the present case there is no
reason to speed the conversion process. Clearly, interrupts could be used with
BUSY tied to the PIO STROBE line.

4)

Read Data from AID Converter

Reading the data Involves a single input operation. We should note that the Analog
Devices AD7570J has an Enable input and tristate outputs so that it could be tied
directly to the microprocessor Data Bus.
The 7570 converter is, of course, underutilized In thiS particular application, particularly since we are interfaCing it to the Z80 processor through a PIO. A simpler 8-bit
AID converter such as the National 5357 device would do the Job at lower cost
this device IS available In an 18-pln package, has a START CONVERSION input, and
provides tristate outputs. It also has output latches and an END OF CONVERSION
output signal.

16-20

5)

Convert Data to Degrees Celsius

Flowchart:

Value = Data rec3ived

from

AID converter

Index == 0
Pornter. "" Start of table

Yes

Index = Index + 1
Pointer = Pointer + 1

Temperature

= Index

The conversion uses a table that contains the largest inUSING A

put value corresponding to a given temperature. The proCALIBRATION
gram searches the table, looking for a value greater than
TABLE
or equal to the value received from the converter. The first
such value it finds corresponds to the required temperature: that IS, if the tenth
entry is the first value larger than or equal to the data, the temperature is 10
degrees. ThiS search method IS inefficient but adequate for the present application.
Note that we must keep the entry number In decimal rather than binary. The instruction sequenCe" ADD A.l . DAA" keeps the index as two decimal digits instead of a binary number. For example, the entry number after 9 (00001001 binary) will be decimal 10 (00010000 BCD) rather than binary ten (00001010). The
reason for thiS IS that we plan to displav the temperature as two deCimal digits and
would have to convert it from binarv to deCimal otherwise.
The table could be obtained by calibration or bv a mathematical approximation.
The calibration method IS simple, since the thermometer must be calibrated anywav The table occupies one memory location for each temperature value to be
displayed. 1
To calibrate the thermometer. you must first adjust the potentiometers to produce
the proper overall range and then determine the converter output values corresponding to specific temperatures.

16-21

6)

Prepare Data for Display

Flowchart:

Get feast significant

digit and sat
output to LEOs

Get most
significant digit

Set output to LEOs

The least significant digit is masked off. We set the bit that
turns on the displays. The result IS saved In Register E.

BLANKING
A LEADING
ZERO

The only difference for the most significant digit IS that a leading zero IS blanked (i.e.. the displays show "blank T rather
than "OT for 7C). This simplY involves not setting the bit that turns on the displays if the digit is zero. The result IS saved In Register D.

16-22

7)

Display Temperature for Six Seconds

Flowchart:

Count

= TSAMP

Send most
significant digit
to left displav

Wait 2 ms

Send least
SIgnificant digit
to right display

Walt 2ms

Count

= Count - 1

Each display is pulsed often enougn so that It appears to be lit continuously. If

TPULS were made longer (say 50 msl. the displays would appear to flash on and
off.
The program uses a 16-bit counter to count the time between temperature samples. The
has instructions to increment or decrement 16-blt register pairs or index registers. However. these Instructions do not affect the flags. so there IS no way
to directly determine when the counter reaches zero. So we make this determination by logically GRing the eight most significant and the eight least significant bits
of the counter. If that result is zero. the 16-blt counter IS zero.

zao

16-23

PROGRAM NAME: THERMOMETER

DATE OF PROGRAM: 10/20/78
:PROGRAMMER: L.ANCE A. LEVENTHAL
:PROGRAM MEMORY REQUIREMENTS: 1548YTES
:RAM REQUIREMENTS. NONE
:1/0 REQUIREMENTS: 1 INPUT PORT, 1 OUTPUT PORT (1 Z80 PIO)
:THIS PROGRAM IS A DIGITAL THERMOMETER THAT ACCEPTS INPUTS FROM
AN A/D CONVERTER ATIACHED TO A THERMISTOR. CONVERTS THE INPUT
, TO DEGREES CELSIUS, AND DISPLAYS THE RESULTS ON TWO
SEVEN-SEGMENT LED DISPLAYS
:A/D CONVERTER
THEA/D CONVERTER IS AN .ANALOG DEVICES 7570J MONOLITHIC CONVERTER
WHICH PRODUCES AN 8-BIT OUTPUT
.THE CONVERSION PROCESS IS STARTED BY A PULSE ON THE START
CONVERSION LINE (BIT 7 OF PIO PORT B)
:THE CONVERSION IS COMPLETED IN 50 MICROSECONDS AND THE
, DIGITAL DATA IS LATCHED
:DISPLAYS
.TWO SEVEN-SEGMENT LED DISPLAYS ARE USED WITH SEPARATE DECODERS
: (7447 OR 7448 DEPENDING ON THE TYPE OF DISPLAY)
;THE DECODER DATA INPUTS ARE CONNECTED TO BITS 0 TO 3 OF
PIO PORT B
:BIT 4 OF PIO PORT B IS USED TO ACTIVATE THE LED DISPLAYS
(BIT 4 IS 1 TO SEND DATA TO LEDS)
:BIT 5 OF PIO PORT B IS USED TO SELECT WHICH LED IS BEING
USED (BIT 5 IS 1 IF THE LEADING DISPLAY IS BEING USED.
, 0 IF THE TRAILING DISPL.AY IS BEING USED)
:METHOD
:STEP 1 - INITIALIZATION
THE MEMORY STACK (USED FOR SUBROUTINE RETURN ADDRESSES) IS
INITIALIZED
:STEP 2 - PULSE START CONVERSION LINE
, THE A/D CONVERTER'S START CONVERSION LINE (BIT 7 OF PIO
PORT B) IS PULSED
;STEP 3 - WAIT FOR AID OUTPUT TO SETILE
A WAIT OF 1 MS ALLOWS FOR COMPLETION OF THE CONVERSION
;STEP 4 - READ A/D VALUE. CONVERT TO DEGREES CELSIUS.
, A TABLE IS USED FOR CONVERSION IT.'CONTAINS THE MAXIMUM
INPUT VALUE FOR EACH TEMPERATURE READING
;STEP 5 - DISPLAY TEMPERATURE ON LEDS
THE TEMPERATURE IS DISPLAYED ON THE LEOS FOR SIX SECONDS
BEFORE ANOTHER CONVERSION IS PERFORMED

.THERMOMETER VARIABLE DEFINITIONS

:MEMORY SYSTEM CONSTANTS

16-24

BEGIN
LASTM

EQU
EQU

50H
1000H

:STARTING ADDRESS OF MAIN PROGRAM

:STARTING ADDRESS FOR RAM STACK

:1/0 UNITS AND PIO ADDRESSES

PIODRA
PIOCRA
PIODRB
PIOCRB

EQU
EQU
EQU
EQU

OEOH
OEZH
OE1H
OE3H

:INPUT PIO FOR CONVERTER

EQU
EQU
EQU
EQU
EQU
EQU

:BIT POSITION TO SEND DATA TOLEDS

;BIT POSITION TO SELECT LEADING DISPLAY
;COUNT NEEDED TO GIVE 1 MS DELAY
;OUTPUT TO BRING START CONVERSION HIGH
:DISPLAY PULSE LENGTH IN MS
;TSAMP IS THE NUMBER ,OF TIMES THE
;DISPLAYS ARE PULSED. IN A
:TEMPERATURE SAMPLING PERIOD. THE
:LENGTH OF A SAMPLING PERIOD IS THUS
;Z*TPULS*TSAMP MILLISECONDS.THE FACTOR
:OF Z*TPULS IS INTRODUCED BY THE FACT
.THAT EACH OF Z DISPLAYS IS PULSED FOR
. TPULS MS

ORG

:OUTPUT PIO FOR DISPLAYS

:DEFINITIONS
LEDON
LEDSL
MSCNT
STCON
TPULS
TSAMP

5
OF9H
10000000B
Z
1500

:RESET ROUTINE TO REACH THERMOMETER PROGRAM

JP

BEGIN

:FIND THERMOMETER PROGRAM

:INITIALIZATION OF THERMOMETER PROGRAM

ORG
LD
OUT
LD
OUT
LD

BEGIN
A.Ol00llllB
(PIOCRA).A
A.OOOOllll B
(PIOCRB).A
SP.LASTM

:MAKE PIO PORT A INPUT

:MAKE PIO PORT B OUTPUT
:PUT STACK AT END OF RAM

:PULSE START CONVERSION LINE

START:

LD
OUT
SUB
OUT

A.STCON
(PIODRBl.A
A
(PIODRBl.A

;SEND START CONVERSION HIGH

:SEND START CONVERSION LOW

16-25

:DELAY 1 MS FOR CONVERSION

LD
CALL

A.l
DELAY

:CONVERSION DELAY TIME IN MS

:WAIT FOR CONVERSION

:READ DIGITAL DATA FROM CONVERTER

IN

A. (PIODRA)

:GET DATA FROM AID CONVERTER

:CONVERT AID DATA TO 2 BCD DIGITS

CALL

CONVR

:CONVERT DATA TO BCD

:GET LEAST SIGNIFICANT DIGIT

LD
AND
SET
LD

B.A
OFH
LEDON.A
E.A

:SAVE BCD DIGITS

:MASK OFF LSD
:SET OUTPUT TO LEDS
:SAVE LSD IN REGISTER E

:GET MOST SIGNIFICANT DIGIT. BLANK LEADING ZERO

LD
RRCA
RRCA
RRCA
RRCA
AND

JR
SET
SET
SVMSD: LD

A.B

;RESTORE BCD DIGITS

:SHIFT MSD

OFH
Z.SVMSD
LEDON.A
LEDSL.A
D.A

:MASK OFF MSD

:DON'T TURN DISPLAY ON IF VALUE ZERO
:SET OUTPUT TO LEOS
:SELECT LEADING DISPLAY
:SAVE MSD IN REGISTER 0

:PULSE THE LED DISPLAYS

DSPLY

LD
LD
OUT
LD
CALL
OUT
LD
CALL
DEC
LD
OR

JR
JP

C.PIODRB
HL.TSAMP

(CLD
A.TPULS
DELAY
(C).E
A.TPULS
DELAY
HL
A.H
L
NZ.DSPLY
START

:GET OUTPUT PORT ADDRESS

:GET 16-BIT PULSE COUNTER
:OUTPUT LEADING DIGIT TO DISPLAY
:DELAY DISPLAY PULSE LENGTH
:OUTPUT TRAILING DIGIT TO DISPLAY
:DELAY DISPLAY PULSE LENGTH
:COUNT DOWN 16-BIT COUNTER
:REMEMBER DEC HL DOES NOT SET Z FLAG
:CONTINUE PULSING DISPLAYS
:GO SAMPLE TEMPERATURE AGAIN

:SUBROUTINE DELAY WAITS FOR THE NUMBER OF MILLISECONDS SPECIFIED

IN REGISTER A

16-26

DELAY:
DLY1<
WTLP:

EXX
LD
DEC

C,MSCNT

JR

NZ.wTLP

DEC

JR

NZ.DLY1

EXX
RET

;SAVE USER REGISTERS

;LOAD REGISTER C FOR 1 MS DELAY
;WAIT 1 MS
:COUNT DOWN NUMBER OF MS
:RESTORE USER REGISTERS

;SUBROUTINE CONVR CONVERTS INPUT FROM AID CONVERTER TO DEGREES

, CELSIUS BY USING A TABLE. INPUT DATA IS IN THE ACCUMULATOR.
, RESULT IS 2 BCD DIGITS IN THE ACCUMULATOR
:REGISTERS USED: A,B,C,H.L

CONVR;

CHVAL:

LD

HL,DEGTB

LD
LD
LD
CP
LD
RET
ADD
DAA
LD
INC

B,A
C.O
A.(HL)
B
A,C
NC

JR

A,l

:GET BASE ADDRESS OF CONVERSION

, TABLE
;SAVE AID INPUT
:START DEGREES AT ZERO
;GET ENTRY FROM TABLE
;IS AID INPUT BELOW ENTRY?
;GET VALUE IN DEGREES CELSIUS
:YES, VALUE FOUND
:NO. ADD 1 TO DEGREES
;KEEP DEGREES IN BCD

C.A
HL
CHVAL

:TABLE DEGTB WAS OBTAINED BY CALIBRATION WITH A KNOWN REFERENCE

;DEGTB CONTAINS THE LARGEST INPUT VALUE THAT CORRESPONDS TO A
PARTICULAR TEMPERATURE READING (I.E., THE FIRST ENTRY IS DECIMAL
58 SO AN INPUT VALUE OF 58 IS THE LARGEST VALUE GIVING A ZERO
, TEMPERATURE READING - VALUES BELOW ZERO ARE DISPLAYED AS ZERO

16-27

DEGTB:

DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
DEFB
END

5B
61
63
66
69
71
74
77
80
84
87
90
93
97
101
104
108
112
116
120
124
12B
132
136
141
145
149
154
158
163
167
172
177
181
186
191
195
200
204
209
214
218
223
227
232
236
241
245
249
253
255

16-28

References
1. A method that uses far less memory IS described In T. A. Selm. "Numencallnterpolatlon for Microprocessor-based Systems:' Computer Design. February 1978. pp. 111-

116.

--

See also:
2. Auslander. D. M. et al., "Direct Digital Process Control: Practice and Algonthms for
Microprocessor ApplicatIOns:' Proceedings of the IEEE. February 1978. pp. 199-208

----

3. Bernstein. N.. "What to Look for

ary 19. 1978. pp. 13-119

4. Bibbero. R. J.. Microprocessors

In

In

Analog Input/Output Boards:' Electronics. JanuInstruments and Control, Wiley. New York. 1977

5. Burton. D. P. and A. L. Dexter. Microprocessor Systems Handbook. Analog Devices.

Inc.. P.O. Box 796. Norwood. MA. 02062. 1977

6. Finkel, J.. Computer-Aided ExpenmentatlOn. Wiley. New York. 1975

7, Garrett. P H.. Analog Systems for Microprocessors and Minicomputers. Reston Publishing Co.. Reston. VA.. 1978
8. Hnatek. E. R.. A User's Handbook of D/A and A/D Converters. Wiley. New York.

1976
9. Mrozowskl. A.. "Analog Output Chips Shnnk A-D Conversion Software:' Electronics.
June 23. 1977, pp. 130-133

10. The Optoelectronics Data Book. Texas Instruments. Inc.. P,O, Box 5012. Dallas. TX..
1978
11. The Optoelectronic DeSigner's Catalog. Hewlett-Packard Inc.. 1820 Embarcadero
Road. Palo Alto. CA. 94303.1978
12. Peatman. J. B.. Microcomputer-based DeSign. McGraw-HilL New York. 1977
13. Rony. P R. et aL. "Microcomputer InterfaCing: Sample and Hold DeVices:' Computer
Design. December 1977. pp. 106-108
14. Sheingold. D. H. ed.. Analog-Digital Conversion Notes. Analog DeVices. Inc.. P O.
Box 796. Norwood. MA. 02062. 1977

16-29

Index of Instruction Descriptions

ADC
ADC
ADC
ADC
ADC
ADC
ADD
ADD
ADD
ADD
ADD
ADD
ADD
AND
AND
AND
AND
AND
81T
BIT
BIT
BIT

A.data 3-43
A.reg 3-44
A.(HLI 3-45
A.(lX + displ 3-45
A.(ly + displ 3-45
HL.rp 3-46
Adata 3-47
A.reg 3-48
A.(HLI 3-49
A.(lX + displ 3-49
A(lY + displ 3-49
HL.rp 3-50
XY.rp 3-51
data 3-52
reg 3-53
IHLI 3-54
(IX + disp) 3-54
(lY + displ 3-54

b.reg 3-55
b.(HLI 3-56
b.(lX + displ
b.IIY + disp)

3-56
3-56

CALL label 3-57

CALL conditlon.label 3-58
CCF 3-59
CP data 3-60
CP reg 3-61
CP IHLI 3-62
CP (IX + displ 3-62
CP (lY + displ 3-62
CPD 3-63
CPDR 3-64
CPI 3-65
CPIR 3-66
CPL 3-67
DAA 3-68
DEC reg 3-69
DEC rp 3-70
DEC IX 3-70
DEC IY
DEC (HLI 3-71
DEC (IX + disp) 3-71
DEC !lY + disp) 3-71
DI 3-72
DJNZ disp 3-73
EI 3-73
EX AF.AF
EX DE.HL
EX (SPI.HL
EX (SPl.IX
EX (SPl.IY
EXX 3-78
HALT

3-75
3-76
3-77
3-77
3-77

3-79

1M 0 3-80
1M 1 3-80
1M 2 3-80
IN A.(portl 3-81
INC reg 3-82
INC rp 3-83
INC IX 3-83
INC IY 3-83
INC IHLI 3-84
INC !IX + disp) 3-84
INC !lY + displ 3-84
IND 3-85
INDR 3-85
INI 3-86
INIR 3-86
IN reg.(CI 3-87
JP
JP
JP
JP
JP
JR
JR
JR
JR
JR

label 3-88
condition.label
(HLI 3-90
!IX) 3-90
!lY) 3-90
C.disp 3-91
disp 3-92
NC.disp 3-93
NZ.disp 3-93
Z.disp 3-94

LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD
LD

A.I 3-94
A.R 3-94
A.(addr) 3-95
Alrp) 3-96
ds!.src 3-97
HL,(addr) 3-98
rp.(addrl 3-98
IX.laddri 3-98
IY.(addri 3-98
I.A 3-99
R.A 3-99
reg.data 3-100
rp.data 3-101
IX.data 3-101
IY.data 3-101
reg.IHLI 3-102
reg.OX + disp) 3-102
reg.OY + disp) 3-102
SP.HL 3-103
SP.IX 3-103
SP.IY 3-103
laddrl.A 3-104
laddrl.HL 3-105
(addrl.rp 3-105
laddrl.xy 3-105
(HLI.data 3-107
(IX + displ.data 3-107
!lY + displ.data 3-107

xv

3-89

Index of Instruction Descriptions (Continued)

LD IHLI.reg 3-108
LD (IX + displ.reg 3-108
LD (lY + displ.reg 3-108
LD (rpl.A 3-109
LDD 3-110
LDDR 3-111
LDI 3-112
LDIR 3-113
NEG
NOP

3-113
3-114

OR data 3-115
OR reg 3-116
OR (HLI 3-117
OR (IX + displ 3-117
OR (lY + displ 3-117
OUT (Cl.reg 3-118
OUTD 3-119
OTDR 3-119
OUTI 3-120
OTIR 3-120
OUT (portl.A 3- 121
POP rp 3-122
POP IX 3-122
POP IY 3-122
PUSH rp 3-123
PUSH IX 3-123
PUSH IY 3-123
RES b.reg 3-124
RES B.IHLI 3-125
RES b.(lX + displ 3-125
RES b.(ly + displ 3-125
RET 3-126
RET cond 3-127
RETI 3-128
RETN 3-129
RL reg 3-130
RL (HLI 3-131
RL (IX + displ 3-131
RL (lY + displ 3-131
RLA 3-132
RLC reg 3-133
RLC (HLI 3-133
RLC (IX + displ 3-134
RLC (lY + displ 3-134
RLCA 3-135

RLD 3-136
RR reg 3-137
RR (HLI 3-138
RR (IX + displ 3-138
RR (lY + displ 3-13B
RRA 3-139
RRC reg 3-140
RRC IHLI 3-141
RRC (Ix + displ 3-141
RRC (ly + displ 3-141
RRCA 3-142
RRD 3-143
RST n 3-144
SBC A.data 3-145
SBC A.reg 3-146
SBC A.IHLI 3-147
SBC A.(lX + displ 3-147
SBC A(lY + displ 3-147
SBC HLrp 3-14B
SCF 3-149
SET b.reg 3-150
SET b.(HLI 3-151
SETb.(lX+displ 3-151
SET b.(lY + displ 3-151
SLA reg 3-152
SLA IHLI 3-153
SLA (Ix + displ 3-153
SLA (IY + displ 3-153
SRA reg 3-154
SRA (HLI 3-155
SRA (IX + displ 3-155
SRA (lY + displ 3-155
SRL reg 3-156
SRL (HLI 3-157
SRL (Ix + displ 3-157
SRL (ly + displ 3-157
SUB data 3-158
SUB reg 3-159
SUB (HU 3-160
SUB (Ix + dis pi 3-160
SUB (IY + displ 3-160
XOR
XOR
XOR
XOR
XOR

xvi

data
reg
(HLI
(IX +
(IY +

3-161
3-162
3-163
displ 3-163
displ 3-163

Index
Accumulator. using the. 4-2
Add/Subtract flag. 8-7
.Address field. numbers and characters in. 3-172
Algebraic notation. 1-8
Algorithm
multiplication. 8-8
simple sorting. 9-10
Allocating RAM. 2-7
Arithmetic and Logical Expressions. 2-10
ASCII
characters. 2-10
handling data In. 6-1
Assembler. 1-5
arithmetic and logical operations. 3-172
chOOSing an. 1-6
meta-.2-14
mlcro-. 2-14
one-pass. 2-14
reSident. 2-14
two-pass. 2-14
Assembler directive. 2-4
Assembly language
applications. 1-10
fields. 2-1
program. 1-5
BasIc software delav. 11-8
BCD and blnarv. accuracv In. 8-8
Blanking a leading zero. 16-22
Block I/O instruction. 6-6
use of. 11-21
Block. moving data within. 7-8
Block search instructions. 6-6
Block transfer instructions. 8-4
Binary and BCD. accuracy In. 8-8
Binary instructions. 1-1
rounding. 8-24
Binary numbers. doubling and halving. 8-23
Bootstrap loader. 2-15
Bottom-up design. 13-44
BreakpOint. 14-2
insertion of. 14-3
RST as. 14-2
Buffer
double buffering. 12-7
emptYing with Interrupts. 12-19
filling via Interrupts. 12-16
Buffer. emptYing with Interrupts. 12-19
CalibratIOn table. use of. 16-21
Character format. 11-81
Checklist. what to Include In. 14-10
Coding. 13-3
relative Importance of. 13-1
Commenting
examples. 15-4
gUidelines. 15-2
techniques. 2-13

questions for. 15-4

Common-anode or common-cathode displays.
11-43
Compiler. 1-7
cost of. 1-8
Computer program. 1-1
CO NO and ENDC pseudo-operations. 3-174
Control and status Information. 11-57
Control informatIOn. combining. 11-58
Credit verification terminal. structural program
for. 13-38
Cross-assembler. 2-14
Dalsv chain
device operation In. 12-10
Interrupts. advantages and disadvantages.
12-9
PIO Interrupts. 12-9
Data. forming classes of. 14-28
moving Within a block. 7-8
Data flowcharts. 13-19
Debouncing
In software. 11-26
With cross-coupled NAND gates. 11-28
Debugging. 13-3
code conversion program. 14-6
interrupt-driven programs. 14-14
sort program. 14-6
use of test cases from. 14-27
DeCimal
accuracy In binary. 8-4
adiust. 8-7
data or addresses. 2-9
rounding. 8-24
shift Instructions. 8-21
DEFB. DEFL. DEFM. DEFS. DEFW pseudooperaltlons. 3-170. 3-171
Definition list
rules jor. 15-8
typical. 15-9
Definitions. placement of. 2-7
Delay loop constant. 11-10
Delimiters. 2-2
Direct memory access (DMA). 11-5
Disabling Interrupts. 12-25
Displays. common-anode or common-cathode.
11-43
DiVISion algOrithm. 8-12
Documentation. 13-3
of status and control transfer. 11-59
of subroutines. 10-2
package. 15-13
Double buffering. 12-7
8-blt summation. 5-3
8080A unused operation codes. 3- 164
8080A/Z80
assembly level conversion. 3-164

xvii

Index (Continued)
8080A/Z80 (continued)
compatibility features. 3-164
incompatibilities. 3-164
8085/Z80 Incompatibililies. 3-165
ENDC and COND pseudo-operatIOns. 3-174
Error considerations. 13-5
Errors. common. 14-11
Example format. 4-1
Examples. gUidelines for. 4-1
Execution time. saving. 15-15
External references. 2-8
Flowcharting
advantages of. 13-17
credit verification. 13-22
disadvantages of. 13-18
sections. 13-22
switch and light system. 13-19
sWitch-based memorv loader. 13-20
Flowcharts
data. 13-19
hints for use. 15-7
Format. 2-2
FORTRAN. 1-7
Full-duplex. 11-89
General service routines. tasks for. 12-30
Hand assemblv. 1-5
Hand checking questions. 14-11
Handshake. 11-2
Hashing. 9-4
Hexadecimal loader. 1-3
Hexadecimal or octal. 1-3
High-level language
advantages of. 1-9
applications for. 1-10
disadvantages of. 1-9
inefficiency of. 1-8
machine Independence. 1-7
overhead for. 1-9
portability of. 1-8
syntax of. 1-10
unsUitabilitv of. 1-10

disabling. 12-2. 12-25

disadvantages of. 12-2
enabling. 12-2
emptYing a line buffer with. 12-19
handling bv monitors. 12-13
Inputs. 12-2. 12-3
Instruction. 12-3
keyboard. 12-14
modes. 12-4
non-maskable. 12-2. 12-3
On particular microcomputers. 12-13
PIO. 12-6. 12-7
reasoning behind. 12-1
510.12-26.12-10
start bit interrupt. 12-28
systems. characteristics of. 12-1
I/O
and memory. 11-1
categories. 11-1
driver. 11-18
Instruction examples. 11-19
instructions with absolute addreSSing. 11-18
Jumps. Indirect. 9-15
Key closure. waiting for. 11-62
Key table. 16-7
Keyboard errors. correcting. 13-14
Keyboard Interrupt 12-14
Kevboard routine. expanding the. 13-48
Keyboard scan. 11-60
Label field. 2-2
Labeling. rules of. 2-3
Labels
chOice of. 2-3
In lump instructions. 2-2
Language levels
application areas for. 1-10
future trends In. 1- 11
LED control. 11-39
Link editor. 2-15
Linking loaders. 2-15
Loader
bootstrap. 2-15
hexadecimal. 1-3
linking. 2-15
memory. 13-10. 13-28
relocating. 2-15
Local or global variables. 2-13
Location counter. 2-7
LogiC analyzer. 14-9
Important features of. 14-10
Logical and arithmetiC expressions. 2-10

Index registers. use of. 7-7

Information hiding prlclple. 13-29
Initializing RAM. 2-8
Input. factors In. 13-4
Instructions
defining a sequence of. 2-11
faster and slower executing. 3-164
Interfaces. standard. 11-103
Interfacing
high-speed deVices. 11-5
medium-speed devices. 11-2
slow deVices. 11-2
Interrupts

Machine language
applicatIOns for. 1-10
program. 1-2
MACRO and ENDM pseudo-operations. 3-174

xviii

Index (Continued)
Macro-assembler. 2-14
Macros
advantages of. 2-12
disadvantages of. 2-12
Mamtenance and redesign. 13-3
Matrix kevboard. 11-60
Memory dump. 14-7
Memory loader error handling. 13-10
Memory map. typical. 15-8
Meta-assembler. 2-14
Micro-assembler. 2-14
Mnemonics. problems with. 1-4
Modular programmmg
advantages of. 13-26
disadvantages of, 13-27
rules for. 13-30
Modularization
pnnclples of. 13-27
switch and light system. 13-28
switch-based memorv loader. 13-28
verification terminal. 13-28
Multiplication algonthm. 8-8

Processmg. factors m. 13-5

Program design. 13-3
basIc pnnciples of. 13-6
Programming gUidelines. 4-2
Pseudo-operations. 2-4
COND.3-174
DEF8.3-170
DEFL. 3-171
DEFM.3-170
DEFS.3-171
DEFW.3-170
END. 3-172
ENDC.3-174
ENDM.3-174
EQU.3-171
MACRO. 3-174
ORG.3-171
RAM
allocating. 2-7
initializmg. 2-8
Real-time clock. 12-20
frequency of. 12-20
prioritvof. 12-21
synchrOnization with. 12-20
Real time. mamtammg. 12-24
Receive routine. structured. 13-40
Redesign and maintenance. 13-3
Redesign. cost of. 15-14
Re-entrant subroutine. 10-2
References. external. 2-8
Register dumps. 14-4
Register Pair HL. using. 4-2
Relocating loader. 2-15
Relocation. 10-2
Relocation constant. 2-3
Reorganization. major or mmor. 15-14
Resident assembler. 2-14
Restart mstruction. 12-4
Return address. changing the. 12-16
Rollover. 11-69
RST as a breakpomt. 14-2

Names
choice of. 2-6. 15-2
defining. 2-6
use of. 2-6
Number svstems. 2-9
Numbers. self-checkin9. 8-17
Non-maskable interrupt. 12-2. 12-3
Object program. 1-2. 1-5
Octal or hexadeCimal. 1-3
One-pass assembler. 2-14
Operation codes. two-word. 3-164
Operator error connection in memorv loader.
13-10
Operator interaction. 13-6
ORG pseudo-operation. 3-171
Passmg parameters. 10-1
PIO
addresses. 11-11
bidirectional mode. 11-15
control mode. 11-15
daisy cham signals. 12-9
directions m control mode. 11-15
mput mode. 11-15
mterrupts. enabling and disabling. 12-7
modes. 11-15. 11-16
output mode. 11-15
registers and control lines. 11-11
steps in configunng. 11-17
Polling. 12-2. 12-10
Polling interrupt svstems with SIOs. 12-10
Portabil itv. 1-6
Primed registers. saving values in. 12-16
Pnoritv.12-16
Problem definition. 13-3

Searchmg methods. 9-6

Self-checking numbers. 8-17
Self-documentmg programs. rules for. 15-1
Seven-segment representations. 11-45
Sign propagation. 8-25
Simple sortmg algorithm. 9-10
Single-step. 14-1
Single-step mode. limitations of. 14-2
SIO
addresses. 11-89
configuration. example of. 11-100
error status. 11-100
interrupt routine. 12-26
mterrupts.12-10
read and write register. address mg. 11-89

xix

Index (Continued)
rules for, 14-29
sort program, 14-29
special cases. 14-28
Testing aids, 14-27
Testing, structured, 14-28
Thermometer analog hardware, 16-15
Timing incompatibilities, 3-165
Timing Intervals
methods for producing. 11-8
uses of. 11-8
Timing method. choosing a, 11-8
Top-down design
advantages of, 13-44
disadvantages of. 13-44
format for. 13-49
methods. 13-44
of SWitch and light system. 13-45
of switch-based memory loader. 13-46
of verification terminal. 13-47
Transmission errors
correcting. 13,15
redUCing, 11-5
Transparent delay routine, 11-8
TTL encoder. uSing a. 11-34
TTY
Interface. 11-81
receive mode. 11-81
standard TTY, 11-81
transmit mode. 11-86
Two-pass assembler 2-14
Two-word operation codes. 3-164

SIO (continued)
reset. 11-97
special features of. 11-97
Software developement, stages of. 13-1
Software simulator. 14-8
Source program. 1-5
Special Instructions. 4-3
Standard Interfaces. 11-103
Standard program library forms. 15-10
Standard TTY. 11-81
Start bit Interrupt. 12-28
Status and control transfers. documenting.
11-59
Status changes with instruction execution, 3-22
Status information, separating. 11-58
Stopwatch Input procedure. 16-1
Strobe. 11-5
Structures. examples of. 13-33
terminators for. 13-43
Structured kevboard routine, 13-38
Structured program for credit verification
ter mlnal. 13-38
Structured programming
advantages of. 13-35
basic structures of. 13-31
disadvantages of. 13-35
for sWitch-based memorv loader. 13-36
In SWitch and light system. 13-36
rules for, 13-43
when to use. 13-35
Structured receive routine. 13-40
Structed testing. 14-28
Stubs. 13-44
SUbroutine instructions. 10-1
Subroutine library. 10-1
SUbroutines. documenting. 10-2
SWitch and light error handling. 13-7
SWitch and light input. 13-6
SWitch and light outputs. 13-7
SWitch and light system. defining, 13-6
SWitch-based memory loader. defining. 13-8
SWitch bounce. 11-26
Symbol table. 2-6
Svnchronlzlng with I/O devices. 11-57

UART.11-88
Varrables. local or global. 2-13
Vectorrng. 12-2
Verification terminal
defining a. 13-11
error handling. 13-14
Inputs. 13-13
outputs. 13-13
Z80
delay loop constant. 11-10
Index registers. use of. 7-7
Interrupt Inputs. 12-2
Interrupt InstrucliOn, 12-3
Interrupt response. 12-3
I/O instructions, 11-18
non-maskable interrupt. 12-3

Terminators for structures. 13-43

Testi ng. 13-3
arithmetic program. 14-29

xx

About the Author

Lance A. Leventhal is a partner in Emulative Systems Company, a San
Diego-based consulting firm specializing in microprocessors and
microprogramming. He serves as Technical Editor of the Society for Computer Simulation and as a Contributing Editor for Digital Design. He is a national lecturer on microprocessors for the IEEE. the author of five books
and over forty articles on microprocessors, and a regular contributor to
such publications as Simulation, Digital Design, and Kilobaud.
Dr. Leventhal's previous experience includes affiliations with Linkabit Corporation, Intelcom Rad Tech, Naval Electronics Laboratory Center and Harry Diamond Laboratories. He received a B.A. degree from Washington
University in St. Louis, Missouri, and M.S. and Ph.D. degrees from the
University of California at San Diego. He is a member of SCS, ACM, and
IEEE.

OSBORNE/McGraw-Hili GENERAL BOOKS

An Introduction to Microcomputers series
by Adam Osborne
Volume 0 - The Beginner's Book
Volume 1 - Basic Concepts
Volume 2 - Some Real Microprocessors (1_978 ed.)
Volume 3 - Some Real Support Devices (1978 ed.)
Volume 2 1978-1979 Update Series
Volume 3 1978-1979 Update Series
The 8089 I/O Processor Handbook
by Adam Osborne
The 8086 Book
by R. Rector and G. Alexy
8080 Programming for Logic Design
by Adam Osborne
6800 Programming for Logic Design
by Adam Osborne
Z80 Programming for Logic Design
by Adam Osborne
8080A/8085 Assembly Language Programming
by L. Leventhal
6800 Assembly Language Programming
by L. Leventhal
6502 Assembly Language Programming
by L. Leventhal
Z8000 Assembly Language Programming
by L. Leventhal et al.
Running Wild: The Next Industrial Revolution
by Adam Osborne
PET-CBM Personal Computer Guide
by Carroll Donahue and Janice Enger
PET and the IEEE 488 Bus (GPIB)
by E. Fisher and C. W. Jensen

OSBORNE/McGraw-Hili SOFTWARE
Practical Basic Programs
by L. Poole et a!.
Some Common BASIC Programs
by L. Poole and M. Borchers
Payroll with Cost Accounting - CBASIC
by Lon Poole et a!.
Accounts Payable and Accounts Receivable - CBASIC
by Lon Poole et al.
General Ledger - CBASIC
by Lon Poole et al.
Some Common Basic Programs - PET/CBM
edited by Lon Poole et a!.