Logic Gates
Introduction
Logic gates serve as the building blocks to digital logic circuits using combinational logic. We're going to consider the following gates: NOT gates (also called inverters), AND gates, OR gates, NAND gates, NOR gates, XOR gates, and XNOR gates.
We'll also discuss the concept of gate deltay.
NOT gates
NOT gates or inverters have a single bit input and a single bit of output.
This is a diagram of a NOT gate. It is a triangle with a circle on the right. The circle indicates "negation".
The truth table defines the behavior of this gate.
where
x is the input and
z is the output.
AND2 gates
AND2 gates have two bits of input and a single bit of output. The subscript, 2, indicates how many inputs this AND gate has. For example, AND3 gates have 3 inputs.
The output of AND
2 gate is 1 only if both inputs are 1. Otherwise, the output is 0.
The truth table defines the behavior of this gate.
The function implmented by
AND2 gates has interesting properties:
- The function is symmetric. Thus, x * y == y * x. This can be verified by using truth tables. We use * to represent AND2.
- The function is associative. Thus, (x * y) * z == x * (y * z). This can be verified by using truth tables.
Because of these properties, it's easy to define ANDn, which is an n-input AND gate.
ANDn(x1, x2,...,xn) = x1 * x2 * ... * xn
That is, an AND gate with n-inputs is the AND of all the bits. This is not ambiguous because the AND function is associative (all parenthesization of this expression are equivalent).
OR2 gates
OR2 gates have two bits of input and a single bit of output. The subscript, 2, indicates how many inputs this OR gate has. For example, OR3 gates have 3 inputs.
The output of OR
2 gate is 0 only if both inputs are 0. Otherwise, the output is 1.
The truth table defines the behavior of this gate.
The function implemented by
OR2 gates has interesting properties:
- The function is symmetric. Thus, x + y == y + x. This can be verified by using truth tables. We use "+" to represent OR2
- The function is associative. Thus, (x + y) + z == x + (y + z). This can be verified by using truth tables.
Because of these properties, it's easy to define ORn, which is an n-input OR gate.
ORn(x1, x2,...,xn) = x1 + x2 + ... + xn
That is, an AND gate with n-inputs is the AND of all the bits. This is not ambiguous because the AND function is associative (all parenthesization of this expression are equivalent).
NAND2 gates
NAND2 gates have two bits of input and a single bit of output. The subscript, 2, indicates how many inputs this NAND gate has. For example, NAND3 gates have 3 inputs.
NAND
k gates is define unusually. Since NAND
2 is
not associative, the definition is based on AND
2.
In particular
NANDk(x1, x2,...,xn) = NOT( ANDk(x1, x2,...,xn) )
Thus, NAND
k is the negation of AND
k.
The truth table defines the behavior of this gate. It's the negation of
AND2.
The function implemented by
NAND2 gates has interesting properties:
- The function is symmetric. Thus, x NAND y == y NAND x. This can be verified by using truth tables.
- The function is not associative. This can be verified by using truth tables.
Because of these properties, NANDk is defined from ANDk, and not built from NAND2 gates.
NOR2 gates
OR2 gates have two bits of input and a single bit of output. The subscript, 2, indicates how many inputs this OR gate has. For example, NOR3 gates have 3 inputs.
The output of NOR
2 gate is the negation of OR
2.
The truth table defines the behavior of this gate.
The function implmented by
NOR2 gates has interesting properties:
- The function is symmetric. Thus, x NOR y == y NOR x. This can be verified by using truth tables.
- The function is not associative. This can be verified by using truth tables.
Because of these properties, NORk is defined from ORk, and not built from NOR2 gates.
XOR2 gates
XOR2 gates have two bits of input and a single bit of output.
The output of XOR
2 gate is 1 only if the inputs have opposite values. That is, when one input has value 0, and the other has value 1.. Otherwise, the output is 0.
This is called
exclusive-or. The definition of OR
2 is
inclusive-or, where the output is 1 if either input is 1, or if both inputs are 1.
XOR
2 can be defined using AND
2, OR
2, and NOT.
x XOR y == ( x AND (NOT y) ) OR ( (NOT x) AND y ) == x\y + y\x
Here's a diagram of the XOR
2 gate.
If you look carefully at the drawing of the gate, there is a second arc behind the first one near the inputs. Since this second arc is hard to see, it's usually a good idea to write the word "XOR" inside the gate.
The truth table defines the behavior of this gate.
The function implmented by
XOR2 gates has interesting properties:
- The function is symmetric. Thus, x (+) y == y (+) x. This can be verified by using truth tables. (We use (+) to denote logical XOR--ideally, we'd draw it with a + sign inside a circle, but HTML doesn't seem to have a symbol for this).
- The function is associative. Thus, [ x (+) y ] (+) z == x (+) [ y (+) z ]. This can be verified by using truth tables.
Because of these properties, it's easy to define XORn, which is an n-input XOR gate.
XORn(x1, x2,...,xn) = x1 (+) x2 (+) ... (+) xn
That is, an XOR gate with n-inputs is the XOR of all the bits. This is not ambiguous because the XOR function is associative (all parenthesization of this expression are equivalent).
XNOR2 gates
XNOR2 gates have two bits of input and a single bit of output.
The output of XNOR
2 gate is the negation of XOR
2 and has 1 when both inputs are the same.
If you look carefully at the drawing of the gate, there is a second arc behind the first one near the inputs. Since this second arc is hard to see, it's usually a good idea to write the word "XNOR" inside the gate.
The truth table defines the behavior of this gate.
The function implmented by
XNOR2 gates has interesting properties:
- The function is symmetric. Thus, x XNOR y == y XNOR x. This can be verified by using truth tables.
- The function is associative. Thus, (x XNOR y) XNOR z == x XNOR (y XNOR z). This can be verified by using truth tables.
Because of these properties, it's easy to define XNORn, which is an n-input XNOR gate.
XNORn(x1, x2,...,xn) = x1 XNOR x2 XNOR ... XNOR xn
That is, an XNOR gate with n-inputs is the XNOR of all the bits. This is not ambiguous because the XNOR function is associative (all parenthesization of this expression are equivalent).
(Error-checkers! You may wish to verify this, and email me if this is incorrect!).
Building Blocks
We can use logic gates to build circuits. While we've described 6 gates, you can do with only three gates to build all possible circuits: AND2, OR2, and NOT. In fact, you don't even need all three gates. It can be done in two kinds of gates of less. We'll explain in a future section.
These circuits can implement any truth table.
Valid Combinational Circuits
The inverse is not true. Not every circuit that is built from gates corresponds to a truth table. In particular, you must observe the following rules if it's to correspond to a truth table.
- The output of a gate may only be attached to the input of another gate. (Think of this as a directed edge from output to input).
- There must be no cycles in the circuit. Treat the circuit like a directed graph with directed edges defined in the previous item.
- Although the output of a gate may be attached to more than one input, an input may not have two different outputs attached to it (this would create conflicting input signals).
- Each input of a gate must come from either the output of another gate or a source. A source is a source that generates either a 0 or 1.
Gate Delay
Real gates have delay. In other words, if you change the value of the inputs, say from 0 and 0 to 0 and 1, then the output takes some small amount of time before it changes. This delay is calledgate delay.
This delay is due to the fact that information can travel at most, the speed of light, and in reality, the time it takes to do the computation is not infinitely quick.
This delay limits how fast the inputs can change and yet the output have meaningful values. It also allows certain kinds of circuits to be created that don't follow the rules from the previous section. In particular, flip flops (to be discussed later) can be generated from gates that use cycles.
Why Subscripts?
Most books don't distinguish between an AND2 gate and a AND3 gate. They claim an AND gate is an AND gate, regardless of the number of inputs.
While this is true, I subscript it because an
AND2 and an
AND3 do not have the same truth table. In particular, an
AND2 truth table has 4 rows while an
AND3 has 8. While the two truth tables are related, they still define different functions.
Thus, I make the distinction by subscripting the number of inputs.
Summary
Logic gates are the building blocks of combinational logic circuits. You can buy logic gates from electronic hobby places. These gates are primarily for hobbyists. Each chip usually has about 4 logic gates.
Real computers don't use these kinds of gates, because they take far too much space. With VLSI technology, you can cram millions of gates onto a wafer no bigger that your thumbnail.
The behavior of logic gates can be described by truth tables. However, because these gates are "physical", they have some properties not expressed in truth tables. In particular, gate delay describes the amount of time it takes for the output to change when the input changes. This time is not zero, thus, one must wait a short amount of time for the output to take effect.
We'll discuss how to build circuits from these gates in a later set of notes.
20:05:00
