Williams 07 - Book Chapters.pdf - Książki - neo666x

Jim Williams

4. Is Analog Circuit Design Dead?

.............................

.............................*.*.....................................................”..

Rumor has it that analog circuit design is dead. Indeed, it is widely rcported and

accepted that rigor niortis has set in. Precious filters, integrators, and the like seem

to have been buried beneath an avalanche of microprocessors, ROMs, RAMS, and

bits and bytes. As some analog people see it (peering out from behind their barri-

cades), a digital monster has been turned loose, destroying the elegance of contin-

uous functions with a blitzing array of flipping and flopping waveforms. The intro-

duction of a ”computerized” oscil loscope-the most analog of all instruments-

with no knobs would seem to be the coup de gr4ce.

These events have produced some bizarre behavior. It has been kindly suggested,

for instance, that the few remaining analog types be rounded up and protected as an

endangered species. Colleges and universities offer fcw analog design courscs. And

soine localities have defined copies of Korn and Korn publications, the Philbr-ick

Applications Munuul, and the Linear Applicutiorzs Handbook as pornographic

material, to be kept away from engineering students‘ innocent and impressionable

minds. Sadly, a few well-known practitioners of the art are slipping across the

border (James E. Solomon has stated, for example, that *‘allclassical analog tech-

niques are dead”), while more principled ones are simply leaving town.

Can all this be happening? Is it really so? Is analog dead‘?Or has the hysteria oi‘

the moment given rise to exaggeralion and distorted judgment?

lo answer these questions with any degree of intelligence and sensitivity, it is

iiccessary to consult history. And to start this process. we must examine the

patient’s body.

Analog circuit design is described using such terms as subtractor, int.egrator,

differentiator: and summingjunction. These mathematical operations are performed

by that pillar of analoggery, the operational amplifier. The use of an amplifier as a

computing tool is not entirely ohvious and was first investigated before World War

11. Practical “computing amplifiers” found their first real niche inside electronic

arialog computers (as opposed to mechanical analog computers such as the Norden

bombsight or Bush’s Differential Analyzer). which werc developed in the iate 1940s

and 1950s. These machines were, by current stmdards, monstrous assemblages

made up of large numbers of amplifiers that could be programmed to integrate, sum,

differentiate, and perform a host of mathematical opcrations. Individual amplificrs

performed singular functions, but complex operations werc performed when all the

amplifiers were interconnected in any desired configuration.

Thc analog computer’s forte was its ability to model or simulate cvcnts. Analog

compiltcrs did not (lie out because analog simulations are no longer uscful or do not

approximate rruth; rather, the rise of digital machines made it enticingly easy to usc

digital fakery to sirnulute the sinrulalions.

Is Analog Circuit Design Dead?

Figure 4-1.

Some analog

ies are merely

leaving town.

tvP

As digital systems came on line in the late 1950s and early 1960s, a protracted

and brutally partisan dispute (some recall it as more of a war) arose between the

analog and digital camps. Digital methods offered high precision at the cost of

circuit complexity. The analog way achieved sophisticated results at lower accuracy

and with comparatively simple circuit configurations. One good op amp (eight

transistors) could do the work of 100 digitally configured 2N404s. It seemed that

digital circuitry was an accurate but inelegant and overcomplex albatross. Digital

types insisted that analog techniques could never achieve any significantaccuracy,

regardless of how adept they were at modeling and simulating real systems.

This battle was not without its editorializing. One eloquent speaker was George A.

Philbrick, a decided analog man, who wrote in 1963 (in The Lightning Empiricist,

Volume 11, No. 4, October, “Analogs Yesterday, Today, and Tomorrow,” pp. 3-43),

“In modest applications to on-line measurement and data processing, it is quite

generally conceded that the advantage of continuous analog apparatus make it irre-

sistible. This is partly owing to the simplicity and speed which its continuity makes

possible, and partly to the fact that almost every input transducer is also ‘analog’ in

character, that is to say, continuous in excursion and time.”

Philbrick, however, a brilliant man, was aware enough to see that digital had at

least some place in the lab: “Only the most hard-shelled of analog champions would

suggest that all simulative and computational equipment be undiluted by numerical

or logical adjuncts.”

He continued by noting that “some analog men, perhaps overfond and defensive

as regards continuous functions, really believe that analog operations are general-

izations of digital ones, or that conversely digital operations are special cases of

analog ones. What can be done with such people?

“While it is agreed that analog and digital techniques will increasingly cross-

fertilize and interrelate,” Philbrick concluded, “it is predicted that the controversy

between their camps will rage on, good natured but unabated, for years to come in

spite of hybrid attachments.”

Although Philbrick and others were intelligent enough to prevent their analog

passions from obscuring their reasoning powers, they could not possibly see what

was coming in a very few years.

Jim Williams

Figure4-2.

Is this the fate of

oscilloscopes

whose innards

are controlled by

knobs instead of

microchips?

Jack Kilby built his IC in 1958.By the middle 1960s, RTL and DTL were in

common use.

While almost everyone agreed that digital approximations weren’t as elegant as

“the real thing,” they were becoming eminently workable, increasingly inexpensive,

and physically more compactable. With their computing business slipping away,

the analog people pulled their amplifiers out of computers, threw the racks away,

and scurried into the measurement and control business. (For a nostalgic, if not

tearful, look at analog computers at the zenith of their glory, read A Palimpsest on

the Electronic Analog Art, edited by Henry M. Paynter.)

If you have read thoughtfully to this point, it should be obvious that analog is

not dead, rather just badly shaken and overshadowed in the aftermath of the war.

Although measurement and control are certainly still around, the really glamorous

and publicized territory has been staked out by the digital troops for some time.

Hard-core guerrilla resistance to this state of affairs, while heroic, is guaranteed

suicide. To stay alive, and even prosper, calls for skillful bargaining based on thor-

ough analysis of the competition’s need.

The understanding that analog is not dead lies in two key observations. First, to

do any useful work, the digital world requires information to perform its operations

upon. The information must come from something loosely referred to as “the real

world.” Deleting quantum mechanics, the “real world” is analog. Supermarket

scales, automobile engines, blast furnaces, and the human body are all examples of

systems that furnish the analog information that the silicon abacus requires to jus-

Williams 07 - Book Chapters.pdf

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: