PRESENTATION TOPIC

IN

DIGITAL DESIGN (20EC21002)

By

Name of the student -HT. No.

Name of the student -HT. No.

Name of the student -HT. No.

Department of Computer Science and Engineering

Geethanjali College of Engineering and Technology

2023 -2024

1
 Include the Content from here

(Use Times New Roman Font 16 for Title in bold, 14 for Heading in bold and 12 for content.

Give 1.5 Line spacing)

2
SYNTHESIS OF THRESHOLD NETWORKS

IN

DIGITAL DESIGN (20EC21002)

By

Name of the student -HT. No.

Name of the student -HT. No.

Name of the student -HT. No.

Department of Computer Science and Engineering

Geethanjali College of Engineering and Technology

2023 -2024

3
 Introduction
Threshold elements are another type of switching elements. Logic Circuits constructed using threshold
elements usually consists of fewer elements and simpler interconnections compared to conventional
gates. The conventional gates logic is specified by Boolean algebra whereas the logic with threshold
gates are specified by mathematical equations. All basic logical gates and universal gates can be
implemented using a Threshold gate (XOR gate cannot be implemented using a single Threshold gate).
Also, some simpler Boolean logics can be implemented using a single Threshold gate.

Threshold Elements:

A threshold element has n two-valued inputs x1, x2,..., xn and a single two valued output y. Its internal
parameters are a threshold T whereas each weight wi associated with a particular input variable xi.
These values of T and xi may be any real, finite, positive or negative numbers. The relation of a
threshold element is defined as follows:

where the sum and product operations are conventional arithmetic ones. The sum is called
the weighted sum of the element.

Threshold element Symbol

4
 Example

Write the input output relation of the Threshold gate given below and obtain the switching funtion

for thesame.

Ans: The inputs are the three x1, x2 and x3 with multiplication factors -1, 2 and 1. The Threshold value T is

½. The output Y is calculated as for the following table.

For the output Y, 1 is entered if the weighted sum is greater than ½. Otherwise, 0 is entered.

Y = f ( x1, x2, x3) = ∑(1,2,6,7) = x1'.x3 + x2 ( after simplification).

Synthesis of Threshold Networks

The methods for the identification and realization of threshold functions and the synthesis of networks of
threshold elements (called as threshold networks) is done in the following ways:

Unate Function:
5
A function f(x1, x2, .. xn) is said to be positive in a variable x, if there exists a disjunctive or conjunctive
expression for the function in which xi appears only in uncompleted form. A function f(x1, x2, ..,xn) is said
to be negative in xi if there exists a disjunctive or conjunctive expression for f in which xi appears only in
complemented form. If f is either positive or negative in xi, then it is said to be unate in xi. No variable
appears in both its complemented and uncompleted form.

Ex 1: The function f= x1x2' + x2x3' is positive in x1 and negative in x3 but is not unate in x2. If a function
f(x1, x2, .., xn) is unate in each of its variables, then it is called Unate function.

Ex 2: The function f = x1'x2 + x1.x2.x3' is unate because by simplification, f = x1 '.x2 + x2.x3' has no variable
in both its complemented and uncompleted form.

Ex 3: The function f = x1. x2' + x1'.x2 is clearly not unate.

If f (x1, x2, …, xn) is positive in xi, then it can be expressed as

If f (x1, x2, .., xn) is negative in xi, then it can be expressed as

Hence, the existence of two such functions, g1 and h1 ( g2 and h2), is a necessary and sufficient condition
for f to be positive (negative) in xi.

Geometric representation of Unate functions:

Unate functions can be represented in a better way as a cube. For an n variable function, an n cube with 2 n
vertices is used, each vertices representing a min term. A line is drawn between every pair of vertices that

6
differ in just one variable.If the function assumes 1 value, that vertices is called true vertices and if the
function assumes 0 value, that vertices is called false vertices.

Ex: Function f = x'.y' + x.z, then the Geometric representation is as follows:

It is a three variable function. Hence, 3 dimensional cube is required. The bolder lines connecting the two
pairs of true vertices i.e. the pair (1,1,1) and (1,0,1) and the pair (0,0,1) and (0,0,0) represent the cubes xz
and x'y' respectively.

A Three-cube representing f=x'y' + xz.

Linear Seperability

If an n cube representation for threshold function with vertices, with a linear equation w1 x1 + w2x2 + … +
wnxn = T correspond to an (n-1) dimensional hyper plane that cuts the n-cube, then

f=0 when w1x1 + w2x2 + … + wnxn < T and

f=1 when w1x1 + w2x2 + … + wnxn >= T

and the hyper plane separates the true vertices from the false ones.

A switching function whose true vertices can be separated by a linear equation from its false ones is called
a linearly separable function and the functional property that makes such a separation is called a linearly
separable function.

7
 Identification and realization of Threshold Functions

To find out whether a given function is a Threshold function and if it is, to calculate the values of the
weights and Threshold value:

a. Test for the Unateness of the given function. Get the minimal form of the given
switching function (both its complement and non-complement of any variable does not exist).

Ex: Given function f = x1.x2.x3 '.x4 + x2.x3'.x4' simplifying it, f = x1.x2.x3' + x2.x3'.x4 '

This is a unate function, as no variable exists in both complement and non-complement form.

b. Convert the given function into another function which has all variables in non-
complement form only. ϕ = x1.x2.x3 + x2.x3.x4

c. Find out all minimal True and maximal false vertices of ϕ. There are two minimal
vertices (1, 1, 1, 0) and (0, 1, 1, 1). The false vertices are (1, 1, 0, 1), (1, 0, 1, 1) and (0, 1, 1, 0).

d. Check whether the function ϕ is linearly separable and if it so, find an appropriate set of
weights and threshold, which is necessary to determine the coefficients of the separating hyper plane.

In the above example, p = 2 and q = 3, there are six inequalities and these are:

From the above, the following are the constraints that must be observed.

8
Letting w1 = w4 = 1 and w2 = w3 = 2, then T must be smaller than 5 but larger than 4. Selecting T=9/2, then
the weight-threshold vector for ϕ V = { 1, 2, 2, 1; 9/2}.

e. convert this weight-threshold vector to one that corresponds to the original function f.
Then the weight-threshold vector will be V = { 1,2,-2,-1; 9/2 -2-1} = { 1,2,-2,-1; 3/2}

Map Based Synthesis of Two Level Threshold Networks

If a given switching function cannot be realized with a single Threshold function, then this non-threshold
function will be decomposed into two or more factors of which each is a Threshold function.

For a function of 3 or 4 variables, the identification problem is solved by detecting certain which are
admissible for Threshold functions, as given below:

If the 1-cells of the given function follow any of the above patterns, then that function can be realized using
a Threshold element. Otherwise, the 1-cells pattern is divided into two admissible patterns.

Ex: Let f (x1, x2, x3, x4) = sigma (2, 3, 6, 7, 10, 12, 14, 15)

The map for this function exhibiting two admissible patterns is

9
The threshold elements for realizing each of the admissible patterns are as below, as per the realization of
the Threshold function described above.

The Threshold logic realization for the function f is

The weight of g in the second element is determined by computing the minimal weighted sum that can
occur in the second element when g has the value 1. Since f must have the value 1 whenever g does, this
minimal weighted sum must be larger than the threshold of the second element. In this case, the minimal
weighted sum is wg, and it occurs when x1 = x2 = 0 and x3 = x4 = 1. Clearly, wg must be larger than 5/2
and, therefore, the value w g = 3 has been selected.

10