Digital Systems - Chapter13.pdf - po ang - Januszek66

CHAPTER 13

PROGRAMMABLE

LOGIC DEVICE

ARCHITECTURES* †

■

OUTLINE

13-1

Digital Systems Family Tree

13-5

The Altera EPM7128S

CPLD

13-2

Fundamentals of PLD

Circuitry

13-6

The Altera FLEX10K

Family

13-3

PLD Architectures

13-7

The Altera Cyclone Family

13-4

The GAL 16V8 (Generic

Array Logic)

*Diagrams of the GAL 16V8 device presented in this chapter have been reproduced through the cour-

tesy of Lattice Semiconductor Corporation, Hillsboro, Oregon.

† Diagrams of the MAX7000S and FLEX10K family devices presented in this chapter have been repro-

duced through the courtesy of Altera Corporation, San Jose, California.

■ OBJECTIVES

Upon completion of this chapter, you will be able to:

■

Describe the different categories of digital system devices.

Describe the different types of PLDs.

■

Interpret PLD data book information.

■

Define PLD terminology.

■

Compare the different programming technologies used in PLDs.

■

Compare the architectures of different types of PLDs.

■

Compare the features of the Altera MAX7000S and FLEX10K families

of PLDs.

■

■ INTRODUCTION

Throughout the chapters of this book you have been introduced to a wide

variety of digital circuits. You now know how the building blocks of digital

systems work and can combine them to solve a wide variety of digital

problems. More complicated digital systems, such as microcomputers and

digital signal processors, have also been briefly described. The defining

difference between microcomputer/DSP systems and other digital systems

is that the former follow a programmed sequence of instructions that the

designer specifies. Many applications require faster response than a

microcomputer/DSP architecture can accommodate and in these cases, a

conventional digital circuit must be used. In today’s rapidly advancing

technology market, most conventional digital systems are not being

implemented with standard logic device chips containing only simple gates

or MSI-type functions. Instead, programmable logic devices, which contain

the circuitry necessary to create logic functions, are being used to imple-

ment digital systems. These devices are not programmed with a list of

instructions, like a computer or DSP. Instead, their internal hardware is

configured by electronically connecting and disconnecting points in the

circuit.

Why have PLDs taken over so much of the market? With programmable

devices, the same functionality can be obtained with one IC rather than

using several individual logic chips. This characteristic means less board

space, less power required, greater reliability, less inventory, and overall

lower cost in manufacturing.

In the previous chapters you have become familiar with the process of

programming a PLD using either AHDL or VHDL. At the same time, you

have learned about all the building blocks of digital systems. The PLD

implementations of digital circuits up to this point have been presented as

869

870

C HAPTER 13/ P ROGRAMMABLE L OGIC D EVICE A RCHITECTURES

a “black box.” We have not been concerned with what was going on inside

the PLD to make it work. Now that you understand all the circuitry inside

the black box, it is time to turn the lights on in there and look at how it

works. This will allow you to make the best decisions when selecting and

applying a PLD to solve a problem. This chapter will take a look at the

various types of hardware available to design digital systems. We will then

introduce you to the architectures of various families of PLDs.

13-1

DIGITAL SYSTEMS FAMILY TREE

While the major goal of this chapter is to investigate PLD architectures, it is

also useful to look at the various hardware choices available to digital system

designers because it should give us a little better perception of today’s digital

hardware alternatives. The desired circuit functionality can generally be

achieved by using several different types of digital hardware. Throughout this

book, we have described both standard logic devices as well as how program-

mable logic devices can be used to create the same functional blocks.

Microcomputers and DSP systems can also often be applied with the neces-

sary sequence of instructions (i.e., the application’s program) to produce the

desired circuit function. The design engineering decisions must take into ac-

count many factors, including the necessary speed of operation for the circuit,

cost of manufacturing, system power consumption, system size, amount of

time available to design the product, etc. In fact, most complex digital designs

include a mix of different hardware categories. Many trade-offs between the

various types of hardware have to be weighed to design a digital system.

A digital system family tree (see Figure 13-1) showing most of the hard-

ware choices that are currently available can be useful in sorting out the

many categories of digital devices. The graphical representation in the figure

does not show all the details—some of the more complex device types have

many additional subcategories, and older, obsolete device types have been

omitted for clarity. The major digital system categories include standard

logic, application-specific integrated circuits (ASICs) and microprocessor/

digital signal processing (DSP) devices.

Digital

systems

Standard

logic

Microprocessors

and DSP

ASICs

Gate

arrays

Standard

cell

Full

custom

TTL

CMOS

ECL

PLDs

SPLDs

CPLDs

HCPLDs

FPGAs

Fuse

EPROM

EEPROM

EPROM

EEPROM

Flash

SRAM

Flash

Antifuse

FIGURE 13-1

Digital system family tree.

871

S ECTION 13-1/ D IGITAL S YSTEMS F AMILY T REE

The first category of standard logic devices refers to the basic functional

digital components (gates, flip-flops, decoders, multiplexers, registers, coun-

ters, etc.) that are available as SSI and MSI chips. These devices have been used

for many years (some more than 30 years) to design complex digital systems.

An obvious drawback is that the system may literally consist of hundreds of

such chips. These inexpensive devices can still be useful if our design is not

very complex. As discussed in Chapter 8, there are three major families of stan-

dard logic devices: TTL, CMOS, and ECL. TTL is a mature technology consist-

ing of numerous subfamilies that have been developed over many years of use.

Very few new designs apply TTL logic, but many, many digital systems still con-

tain TTL devices. CMOS is the most popular standard logic device family today,

primarily due to its low power consumption. ECL technology, of course, is ap-

plied for higher-speed designs. Standard logic devices are still available to the

digital designer, but if the application is very complex, a lot of SSI/MSI chips

will be needed. That solution is not very attractive for our design needs today.

The microprocessor/digital signal processing (DSP) category is a much dif-

ferent approach to digital system design. These devices actually contain the

various types of functional blocks that have been discussed throughout this

text. With microcomputer/DSP systems, devices can be controlled electroni-

cally, and data can be manipulated by executing a program of instructions that

has been written for the application. A great deal of flexibility can be

achieved with microcomputer/DSP systems because all you have to do is

change the program. The major downfall with this digital system category is

speed. Using a hardware solution for your digital system design is always faster

than a software solution.

The third major digital system category is called application-specific

integrated circuits (ASICs) . This broad category represents the modern

hardware design solution for digital systems. As the acronym implies, an in-

tegrated circuit is designed to implement a specific desired application.

Four subcategories of ASIC devices are available to create digital systems:

programmable logic devices, gate arrays, standard-cell, and full-custom.

Programmable logic devices (PLDs) , sometimes referred to as field-

programmable logic devices (FPLDs), can be custom-configured to create any

desired digital circuit, from simple logic gates to complex digital systems.

Many examples of PLD designs have been given in earlier chapters. This ASIC

choice for the designer is very different from the other three subcategories.

With a relatively small capital investment, any company can purchase the nec-

essary development software and hardware to program PLDs for their digital

designs. On the other hand, to obtain a gate array, standard-cell or full-custom

ASIC requires that most companies contract with an IC foundry to fabricate

the desired IC chip. This option can be extremely expensive and usually re-

quires that your company purchase a large volume of parts to be cost effective.

Gate arrays are ULSI circuits that offer hundreds of thousands of gates. The

desired logic functions are created by the interconnections of these prefabri-

cated gates. A custom-designed mask for the specific application determines

the gate interconnections, much like the stored data in a mask-programmed

ROM. For this reason, they are often referred to as mask-programmed gate

arrays (MPGAs). Individually, these devices are less expensive than PLDs of

comparable gate count, but the custom programming process by the chip

manufacturer is very expensive and requires a great deal of lead time.

Standard-cell ASICs use predefined logic function building blocks called

cells to create the desired digital system. The IC layout of each cell has been

designed previously, and a library of available cells is stored in a computer

database. The needed cells are laid out for the desired application, and

the interconnections between the cells are determined. Design costs for

872

C HAPTER 13/ P ROGRAMMABLE L OGIC D EVICE A RCHITECTURES

standard-cell ASICs are even higher than for MPGAs because all IC fabrica-

tion masks that define the components and interconnections must be custom

designed. Greater lead time is also needed for the creation of the additional

masks. Standard cells do have a significant advantage over gate arrays. The

cell-based functions have been designed to be much smaller than equivalent

functions in gate arrays, which allows for generally higher-speed operation

and cheaper manufacturing costs.

Full-custom ASICs are considered the ultimate ASIC choice. As the name

implies, all components (transistors, resistors, and capacitors) and the inter-

connections between them are custom-designed by the IC designer. This

design effort requires a significant amount of time and expense, but it can

result in ICs that can operate at the highest possible speed and require the

smallest die (individual IC chip) area. Smaller IC die sizes allow for many

more die to fit on a silicon wafer, which significantly lowers the manufactur-

ing cost for each IC.

More on PLDs

This chapter is mainly about PLDs, so we will look a little further down that

branch of the family tree. The development of PLD technology has advanced

continuously since the first PLDs appeared more than 30 years ago. The early

devices contained the equivalent of a few hundred gates, and now we have

parts available that contain a few million gates. The old devices could han-

dle a few inputs and a few outputs with limited logic capabilities. Now there

are PLDs that can handle hundreds of inputs and outputs. Original devices

could be programmed only once and, if the design changed, the old PLD

would have to be removed from the circuit and a new one, programmed with

the updated design, would have to be inserted in its place. With newer de-

vices, the internal logic design can be changed on the fly, while the chip is

still connected to a printed circuit board in an electronic system.

Generally, PLDs can be described as being one of three different types:

simple programmable logic devices (SPLDs), complex programmable logic

devices (CPLDs) , or field programmable gate arrays (FPGAs) . There are sev-

eral manufacturers with many different families of PLD devices, so there are

many variations in architecture. We will attempt to discuss the general char-

acteristics for each of the types, but be forewarned: the differences are not

always clear-cut. The distinction between CPLDs and FPGAs is often a little

fuzzy, with the manufacturers constantly designing new, improved architec-

tures and frequently muddying the waters for marketing purposes. Together,

CPLDs and FPGAs are often referred to as high-capacity programmable logic

devices (HCPLDs) . The programming technologies for PLD devices are actu-

ally based on the various types of semiconductor memory. As new types of

memory have been developed, the same technology has been applied to the

creation of new types of PLD devices.

The amount of logic resources available is the major distinguishing feature

between SPLDs and HCPLDs. Today, SPLDs are devices that typically contain

the equivalent of 600 or fewer gates, while HCPLDs have thousands and hun-

dreds of thousands of gates available. Internal programmable signal intercon-

nect resources are much more limited with SPLDs. SPLDs are generally much

less complicated and much cheaper than HCPLDs. Many small digital applica-

tions need only the resources of an SPLD. On the other hand, HCPLDs are ca-

pable of providing the circuit resources for complete complex digital systems,

and larger, more sophisticated HCPLD devices are designed every year.

The SPLD classification includes the earliest PLD devices. The amount of

logic resources contained in the early PLDs may be relatively small by today’s

Digital Systems - Chapter13.pdf

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: