Logic Circuit Design

(Adapted from notes of N.J. Rao,IISc, courtesy NPTEL)

Number Systems
The present daytechnology and the way of life require the usage of several number systems.
Usageof decimal numbers starts very early in one’s life. Therefore, when one is
confrontedwith number systems other than decimal, some time during the high-school
years, it
calls for a fundamental change in one’s framework of thinking.
There have been two types of numbering systems in use throughout the world.
One type is symbolic in nature. Most important example of this symbolic numbering
system is the one based on Roman numerals
I = 1, V = 5, X = 10, L = 50, C = 100, D = 500 and M = 1000
IIMVII - 2007
While this system was in use for several centuries in Europe it is completely
superseded by the weighted-position system based on Indian numerals. The Roman
number system is still used in some places like watches and release dates of movies.

The weighted-positional system based on the use of radix 10 is the most commonly
used numbering system in most of the transactions and activities of today’s world.
However, the advent of computers and the convenience of using devices that have
two well defined states brought the binary system, using the radix 2, into extensive
use. The use of binary number system in the field of computers and electronics also
lead to the use of octal (based on radix 8) and hex-decimal system (based on radix
16). The usage of binary numbers at various levels has become so essential that it
is also necessary to have a good understanding of all the binary arithmetic
operations.
Here we explore the weighted-position number systems and conversion from one
system to the other.

Weighted-Position Number System

In a weighted-position numbering system using Indian numerals the value
associated with a digit is dependent on its position. The value of a number is
weighted sum of its digits.
Consider the decimal number 2357. It can be expressed as
2357 = 2 x 103 + 3 x 102 + 5 x 101 + 7 x 100
Each weight is a power of 10 corresponding to the digit’s position. A decimal point
allows negative as well as positive powers of 10 to be used;
526.47 = 5 x 102 +2 x 101 + 6 x 100 + 4 x 10-1 + 7 x 10-2
Here, 10 is called the base or radix of the number system. In a general positional
number system, the radix may be any integer r > 2, and a digit position i has weight
ri. The general form of a number in such a system is
dp-1dp-2, .... d1, d0 .d-1d-2 .... d-n
where there are p digits to the left of the point (called radix point) and n digits to the
right of the point. The value of the number is the sum of each digit multiplied by the
corresponding power of the radix.

Except for possible leading and trailing zeros, the representation of a number in
positional system is unique (00256.230 is the same as 256.23). Obviously the
valuesdi’s can take are limited by the radix value. For example a number like
(356)5, where the suffix 5 represents the radix will be incorrect, as there can not be
a digit like 5 or 6 in a weighted position number system with radix 5.
If the radix point is not shown in the number, then it is assumed to be located near
the last right digit to its immediate right. The symbol used for the radix point is a
point (.)
As much of the present day electronic hardware is dependent on devices that work
reliably in two well defined states, a numbering system using 2 as its radix has
become necessary and popular. With the radix value of 2, the binary number system
will have only two numerals, namely 0 and 1.
Consider the number (N)2= (11100110)2.
It is an eight digit binary number. The binary digits are also known as bits.
Consequently the above number would be referred to as an 8-bit number. Its
decimal value is given by

From here on we consider any number without its radix specifically mentioned, as a
decimal number.
With the radix value of 2, the binary number system requires very long strings of 1s
and 0s to represent a given number. Some of the problems associated with handling
large strings of binary digits may be eased by grouping them into three digits or four
digits. We can use the following groupings.
􀂃Octal (radix 8 to group three binary digits)
􀂃Hexadecimal (radix 16 to group four binary digits)
In the octal number system the digits will have one of the following eight values 0, 1,
2, 3, 4, 5, 6 and 7.
In the hexadecimal system we have one of the sixteen values 0 through 15.
However, the decimal values from 10 to 15 will be represented by alphabet A (=10),
B (=11), C (=12), D (=13), E (=14) and F (=15).
Conversion of a binary number to an octal number or a hexadecimal number is very
simple, as it requires simple grouping of the binary digits into groups of three or
four.
Consider the binary number 11011011. It may be converted into octal or
hexadecimal numbers as
(11011001)2 = (011) (011) (001) = (331)8
= (1101) (1001) = (D9)16
Note that adding a leading zero does not alter the value of the number. Similarly for
grouping the digits in the fractional part of a binary number, trailing zeros may be
added without changing the value of the number.
Number System Conversions
In general, conversion between numbers with different radices cannot be done by
simple substitutions. Such conversions would involve arithmetic operations. Let us
work out procedures for converting a number in any radix to radix 10, and viceversa.
The decimal equivalent value of a number in any radix is given by theformula

wherer is the radix of the number and there are p digits to the left of the radix point
and n digits to the right. Decimal value of the number is determined by converting
each digit of the number to its radix-10 equivalent and expanding the formula using
radix-10 arithmetic.
Some examples are:

This forms the basis for converting a decimal number D to a number with radix r. If
we divide the right hand side of the above formula by r, the remainder will be d 0,
and the quotient will be

Thus, d0 can be computed as the remainder of the long division of D by the radix r.
As the quotient Q has the same form as D, another long division by r will give d 1 as
the remainder. This process can continue to produce all the digits of the number
with radix r.

Consider the following examples.

Representation of Negative Numbers
In our traditional arithmetic we use the “+” sign before a number to indicate it as a
positive number and a “-” sign to indicate it as a negative number. We usually omit the
sign before the number if it is positive. This method of representation of numbers is
called “sign-magnitude” representation. But using “+” and “-” signs on a computer is
not convenient, and it becomes necessary to have some other convention to represent
the signed numbers. We replace “+” sign with “0” and “-” with “1”. These two symbols
already exist in the binary system.
Consider the following examples:
In the sign-magnitude representation of binary numbers the first digit is always treated
separately. Therefore, in working with the signed binary numbers in sign-magnitude
form the leading zeros should not be ignored. However, the leading zeros can be
ignored after the sign bit is separated. For example,
1000101.11 = - 101.11
While the sign-magnitude representation of signed numbers appears to be natural
extension of the traditional arithmetic, the arithmetic operations with signed numbers in
this form are not that very convenient, either for implementation on the computer or for
hardware implementation. There are two other methods of representing signed
numbers.
􀂃Diminished Radix Complement (DRC) or (r-1)-complement
􀂃Radix Complement (RX) or r-complement
When the numbers are in binary form
􀂃Diminished Radix Complement will be known as “one’s-complement”
􀂃Radix complement will be known as “two’s-complement”.
If this representation is extended to the decimal numbers they will be known as 9’scomplement
and 10’s-complement respectively.
One’s Complement Representation
Let A be an n-bit signed binary number in one’s complement form.
The most significant bit represents the sign. If it is a “0” the number is positive and if it
is a “1” the number is negative.
The remaining (n-1) bits represent the magnitude, but not necessarily as a simple
weighted number.
Consider the following one’s complement numbers and their decimal equivalents:
0111111 + 63
0000110 --> + 6
0000000 --> + 0
1111111 --> + 0
1111001 --> - 6
1000000 --> - 63
There are two representations of “0”, namely 000.....0 and 111.....1.
From these illustrations we observe
􀂃If the most significant bit (MSD) is zero the remaining (n-1) bits directly indicate
the magnitude.
􀂃If the MSD is 1, the magnitude of the number is obtained by taking the
complement of all the remaining (n-1) bits.
For example consider one’s complement representation of -6 as given above.
􀂃Leaving the first bit ‘1’ for the sign, the remaining bits 111001 do not directly
represent the magnitude of the number -6.
􀂃Take the complement of 111001, which becomes 000110 to determine the
magnitude.
In the example shown above a 7-bit number can cover the range from +63 to -63. In
general an n-bit number has a range from

with tworepresentations for zero.

The representation also suggests that if A is an integer in one’s complement form, then
one’s complement of A = -A
One’s complement of a number is obtained by merely complementing all the digits.
This relationship can be extended to fractions as well.
For example if A = 0.101 (+0.625)10, then the one’s complement of A is 1.010, which is
one’s complement representation of (-0.625)10. Similarly consider the case of a mixed
number.
A = 010011.0101 (+19.3125)10
One’s complement of A = 101100.1010 (- 19.3125)10
This relationship can be used to determine one’s complement representation of negative
decimal numbers.
Example 1: What is one’s complement binary representation of decimal number -75?
Decimal number 75 requires 7 bits to represent its magnitude in the binary form. One
additional bit is needed to represent the sign. Therefore,
one’s complement representation of 75 = 01001011
one’s complement representation of -75 = 10110100
Two’s Complement Representation
Let A be an n-bit signed binary number in two’s complement form.
􀂃The most significant bit represents the sign. If it is a “0”, the number is positive,
and if it is “1” the number is negative.
􀂃The remaining (n-1) bits represent the magnitude, but not as a simple weighted
number.
Consider the following two’s complement numbers and their decimal equivalents:

There is only one representation of “0”, namely 000....0.

From these illustrations we observe
If most significant bit (MSD) is zero the remaining (n-1) bits directly indicate the
magnitude.
If the MSD is 1, the magnitude of the number is obtained by taking the complement of
all the remaining (n-1) bits and adding a 1.
Consider the two’s complement representation of -6.
􀂃We assume we are representing it as a 7-bit number.
􀂃Leave the sign bit.
􀂃The remaining bits are 111010. These have to be complemented (that is
000101) and a 1 has to be added (that is 000101 + 1 = 000110 = 6).
In the example shown above a 7-bit number can cover the range from +63 to -64. In
general an n-bit number has a range from

with one representationfor zero.

The representation also suggests that if A is an integer in two’s complement form, then
Two’s complement of A = -A
Two’s complement of a number is obtained by complementing all the digits and adding
‘1’ to the LSB.
This relationship can be extended to fractions as well.
If A = 0.101 (+0.625)10, then the two’s complement of A is 1.011, which is two’s
complement representation of (-0.625)10.
Similarly consider the case of a mixed number.
A = 010011.0101 (+19.3125)10
Two’s complement of A = 101100.1011 (- 19.3125)10
This relationship can be used to determine two’s complement representation of negative
decimal numbers.
Example 2: What is two’s complement binary representation of decimal number -75?
Decimal number 75 requires 7 bits to represent its magnitude in the binary form. One
additional bit is needed to represent the sign. Therefore,
Two’s complement representation of 75 = 01001011
Two’s complement representation of -75 = 10110101

CODES: Introduction
When we wish to send information over long distances unambiguously it becomes
necessary to modify (encoding) the information into some form before sending, and
convert (decode) at the receiving end to get back the original information. This
process of encoding and decoding is necessary because the channel through which
the information is sent may distort the transmitted information. Much of the
information is sent as numbers. While these numbers are created using simple
weighted-positional numbering systems, they need to be encoded before
transmission. The modifications to numbers were based on changing the weights,
but predominantly on some form of binary encoding. There are several codes in use
in the context of present day information technology, and more and more new codes
are being generated to meet the new demands.
Coding is the process of altering the characteristics of information to make
it more suitable for intended application
By assigning each item of information a unique combination of 1s and 0s we
transform some given information into binary coded form. The bit combinations are
referred to as “words” or “code words”. In the field of digital systems and computers
different bit combinations have different designations.
Bit - a binary digit 0 or 1
Nibble - a group of four bits
Byte - a group of eight bits
Word - a group of sixteen bits;
a word has two bytes or four nibbles
Sometimes ‘word’ is used to designate a larger group of bits also, for example 32 bit
or 64 bit words.
We need and use coding of information for a variety of reasons
􀂃to increase efficiency of transmission,
􀂃to make it error free,
􀂃to enable us to correct it if errors occurred,
􀂃to inform the sender if an error occurred in the received information etc.
􀂃for security reasons to limit the accessibility of information
􀂃tostandardise a universal code that can be used by all
Coding schemes have to be designed to suit the security requirements and the
complexity of the medium over which information is transmitted.
Decoding is the process of reconstructing source information from the
encoded information. Decoding process can be more complex than coding if we
do not have prior knowledge of coding schemes.
In view of the modern day requirements of efficient, error free and secure
information transmission coding theory is an extremely important subject. However,
at this stage of learning digital systems we confine ourselves to familiarising with a
few commonly used codes and their properties.
We will be mainly concerned with binary codes. In binary coding we use binary digits
or bits (0 and 1) to code the elements of an information set. Let n be the number of
bits in the code word and x be the number of unique words.
If n = 1, then x = 2 (0, 1)
n = 2, then x = 4 (00, 01, 10, 11)
n = 3, then x = 8 (000,001,010 ...111)

From this we can conclude that if we are given elements of information to code into
binary coded format,
where j is the number of bits in a code word.
For example, if we want to code alphanumeric information (26 alphabetic characters
+ 10 decimals digits = 36 elements of information), we require
j> 3.32 log1036
j> 5.16 bits
Since bits are not defined as fractional parts, we take j = 6. In other words a
minimum six-bit code would be required to code 36 alphanumeric elements of
information. However, with a six-bit code only 36 code words are used out of the 64
code words possible.
In this Learning Unit we consider a few commonly used codes including
1. Binary coded decimal codes
2. Unit distance codes
3. Error detection codes
4. Alphanumeric codes
Binary Coded Decimal Codes
The main motivation for binary number system is that there are only two elements in
the binary set, namely 0 and 1. While it is advantageous to perform all
computations on hardware in binary forms, human beings still prefer to work with
decimal numbers. Any electronic system should then be able to accept decimal
numbers, and make its output available in the decimal form.
One method, therefore, would be to
􀂃convert decimal number inputs into binary form
􀂃manipulate these binary numbers as per the required functions, and
􀂃convert the resultant binary numbers into the decimal form
However, this kind of conversion requires more hardware, and in some cases
considerably slows down the system. Faster systems can afford the additional
circuitry, but the delays associated with the conversions would not be acceptable. In
case of smaller systems, the speed may not be the main criterion, but the additional
circuitry may make the system more expensive.
We can solve this problem by encoding decimal numbers as binary strings, and use
them for subsequent manipulations.
There are ten different symbols in the decimal number system: 0, 1, 2, . . ., 9. As
there are ten symbols we require at least four bits to represent them in the binary
form. Such a representation of decimal numbers is called binary coding of decimal
numbers.
We wish to choose codes that have some desirable
properties like
􀂃ease of coding
􀂃ease in arithmetic operations
􀂃minimum use of hardware
􀂃error detection property
􀂃ability to prevent wrong output during transitions
In a weighted code the decimal value of a code is the algebraic sum of the weights
of 1s appearing in the number. Let (A)10be a decimal number encoded in the binary
form as a3a2a1a0. Then
(A)10= w3a3 + w2a2 + w1a1 +w0a0
where w3, w2, w1 and w0 are the weights selected for a given code, and a3,a2,a1and
a0are either 0s or 1s. The more popularly used codes have the weights as
The decimal numbers in these three codes are

In all the cases only ten combinations are utilized to represent the decimal digits.
The remaining six combinations are illegal. However, they may be utilized for error
detection purposes.
Consider, for example, the representation of the decimal number 16.85 in Natural
Binary Coded Decimal code (NBCD)
(16.85)10 = (0001 0110 . 1000 0101)NBCD
1685
There are many possible weights to write a number in BCD code. Some codes have
desirable properties, which make them suitable for specific applications. Two such
desirable properties are:
1. Self-complementing codes
2. Reflective codes
When we perform arithmetic operations, it is often required to take the
“complement” of a given number. If the logical complement of a coded number is
also its arithmetic complement, it will be convenient from hardware point of view. In
aself-complementing coded decimal number, (A)10, if the individual bits of a
number are complemented it will result in (9 - A)10.
Example: Consider the 2421 code.
􀂃The 2421 code of (4)10 is 0100.
􀂃Its complement is 1011 which is 2421 code for (5)10 = (9 - 4)10.
Therefore, 2421 code may be considered as a self-complementing code. A necessary
condition for a self-complimenting code is that the sum of its weights should be 9.
A self-complementing code, which is not weighted, is excess-3 code. It is derived
from 8421 code by adding 0011 to all the 8421 coded numbers.
Another self-complementing code is 631-1 weighted code.
Three self-complementing codes are
A reflective code is characterized by the fact that it is imaged about the centre
entries with one bit changed

The most common application of NBCD codes isin the calculator.

Unit Distance Codes
There are many applications in which it is desirable to have a code in which the
adjacent codes differ only in one bit. Such codes are called Unit distance Codes.
“Gray code” is the most popular example of unit distance code. The 3-bit and 4-bit
Gray codes are

These Gray codes listed here have also the reflective properties.
The most popular use of Gray codes is in the position sensing transducer known as
shaft encoder.
There are two convenient methods to construct Gray code with any number of
desired bits. The first method is based on the fact that Gray code is also a reflective
code.However, this method requires Gray codes with all bit lengths less than ‘n’ also be
generated as a part of generating n-bit Gray code.

The second method allows us toderive an n-bit Gray code word directly from the
corresponding n-bit binary codeword:
􀂃The bits of an n-bit binary code or Gray code words are numbered from right
to left, from 0 to n-1.
􀂃Bit i of a Gray-code word is 0 if bits i and i+1 of the corresponding binary
code word are the same, else bit i is 1. When i+1 = n, bit n of the binary
code word is considered to be 0.
Example: Consider the decimal number 68.
(68)10 = (1000100)2
Binary code: 1 0 0 0 1 0 0
Gray code : 1 1 0 0 1 1 0

Gray to binary
Gray code : 1 1 0 1 1 0 1 0 0 0 1 0 1 1
Binary code: 1 0 0 1 0 0 1 1 1 1 0 0 1 0
Alphanumeric Codes
When information to be encoded includes entities other than numerical values, an
expanded code is required. For example, alphabetic characters (A, B, ....Z) and
special operation symbols like +, -, /, *, (, ) and other special symbols are used in
digital systems. Codes that include alphabetic characters are commonly referred to
as Alphanumeric Codes. However, we require adequate number of bits to encode all
the characters. As there was a need for alphanumeric codes in a wide variety of
applications in the early era of computers, like teletype, punched tape and punched
cards, there has always been a need for evolving a standard for these codes.
Alphanumeric keyboard has become ubiquitous with the popularization of personal
computers and notebook computers. These keyboards use ASCII (American
Standard Code for Information Interchange) code

Al
phanumeric codes like EBCDIC (Extended Binary Coded Decimal Interchange Code)
and 12-bit Hollerith code are in use for some applications. However, ASCII code is
now the standard code for most data communication networks.

Error Detection and Correcting Codes

When data is transmitted in digital form from one place to another through a
transmission channel/medium, some data bits may be lost or modified. This loss of
data integrity occurs due to a variety of electrical phenomena in the transmission
channel. As there are needs to transmit millions of bits per second, the data
integrity should be very high. The error rate cannot be reduced to zero. Then we
would like to ideally have a mechanism of correcting the errors that occur. If this is
not possible or proves to be expensive, we would like to know if an error occurred.
If an occurrence of error is known, appropriate action, like retransmitting the data,
can be taken. One of the methods of improving data integrity is to encode the data
in a suitable manner. This encoding may be done for error correction or merely for
error detection.
A simple process of adding a special code bit to a data word can improve its
integrity. This extra bit will allow detection of a single error in a given code word in
which it is used, and is called the ‘Parity Bit’. This parity bit can be added on an odd
or even basis. The odd or even designation of a code word may be determined by
actual number of 1’s in the data (including the added parity bit) to which the parity
bit is added. For example, the S in ASCII code is
(S) = (1010011)ASCII
S, when coded for odd parity, would be shown as
(S) = (11010011)ASCII with odd parity
In this encoded ‘S’ the number of 1’s is five, which is odd.
When S is encoded for even parity
(S) = (01010011)ASCII with even parity.
In this case the coded word has even number (four) of ones.
Thus the parity encoding scheme is a simple one and requires only one extra bit. If
the system is using even parity and we find odd number of ones in the received data
word we know that an error has occurred. However, this scheme is meaningful only
for single errors. If two bits in a data word were received incorrectly the parity bit
scheme will not detect the faults. Then the question arises as to the level of
improvement in the data integrity if occurrence of only one bit error is detectable.
The improvement in the reliability can be mathematically determined.
Adding a parity bit allows us only to detect the presence of one bit error in a group of
bits. But it does not enable us to exactly locate the bit that changed. Therefore,
addition of one parity bit may be called an error detecting coding scheme. In a
digital system detection of error alone is not sufficient. It has to be corrected as
well. Parity bit scheme can be extended to locate the faulty bit in a block of
information.

Hamming Code

Refer Text pages (Photostat given)

Module II
Logic Functions
Karnaugh Maps