Page 34 - Analog Circuit Design Art, Science, and Personalities

Dan Sheingold

3. It’s an Analog World-Or Is It?

Back in the 1950s. I once heard George Philbrick say, “Digital is a special case of
analog.” He was a passionate believer in the analog nature of the world. (He was
also skeptical about frequency, though he understood transform theory-Laplace,
Fourier, and especially Heaviside-better than most. But that‘s a matter for another
essay.)
Now that we’ve had a few more decades to retlect on nature, to observe conver-
gences between organisms and computer programs, and to see ways of simulating
electrical behavior of organisms with computers (e.g., neural nets), it’s possible to
make some definite statements about what’s analog and what’s digital.
First of all, though, we have to dispose of nonlinearity and dismntinuity in nature
as arguments for digital.
Linearity of real-world phenomena has nothing to do with the analog versus
digital question. The real (analog) world is full of nonlinearities. My employer and
others manufacture a number of purposely, predictably, and accurately nonlinear
devices-for example, devices with natural logarithmic or trigonometric (instead
of linearj responses. They are all unalog devices.
Second, discreteness and discontinuity really have little to do with the analog
versus digital question. You don’t have to go to microscopic phenomena to find
discrete analog devices. My employer also manufactures analog switches and coni-
parators. They are discontinuous (hence discretej devices. The switches are funda-
mental digital to analog converters: the comparators are fundamental analog to
digital converters. But voltage or current, representing digitul yuaniities, operates
the switches; and the outputs of the comparators are voltages, representing the choice
oj‘u digitul I or 0. Thus, these basic data converters arc analog to unalog devices.
Perhaps nature is discrete a1 the limits: current could, in a sense, be counted as a
flow of discrete charge carriers; time could be counted as ticks of a clock. And
noise lihits thc resolution of continuous measurements. which some might use to
argue against the continuous case. But these arguments also work against the dis-
crete cilse. The uncertainty principle says we can’t locate a charge carrier and at the
same time say accurately how fast it’s going. So we measure current as the nwrage
number of charge carriers that tlow in a circuit and call the individual carriers noise.
Similarly, a clock that ticked with every event would be useless because it would
tick irregularly, so again we choose a clock that averages the basic ticks, and call
the basic ticks jitter.
Perhaps it.’s useful to accept the duality of discrete and continuous in the analog
real world, even as most people accept that natural phenomena are both particles
(discretej and waves (continuous).
The iniportuntpoint is thut “digital” is irrelevurzt to all that. Digital in the quan-
titative sense applied to physical phenomena is a human concept; it didn’t cxist

15