Page 30 - Analog Circuit Design Art, Science, and Personalities

George A. Philbrick

Digital computers, however, cannot be used conveniently or efficiently to
obtain answers to all of the problems. In some cases, even they cannot solve the
equations in any reasonable time, and in other cases the problems are not under-
stood well enough for satisfactory mathematical formulation. Under these cir-
cumstances we can often turn to analog, real-time, simulation devices to predict
the behaviour of the system. No engineering computing center is well equipped
without such devices.

One should certainly be happy to settle for this, even though the text continues in
a discussion of other kinds of equipment than analog with which the latter may be
associated. Only the most hard-shelled of analog champions would suggest that
all simulative and computational equipment be undiluted by numerical or logical
adjuncts. Certainly many of the best known individuals and organizations in the
analog field are now willing and able to talk about hybrids. This term, by the way,
is too broad to have much meaning at this stage of the game. Is an analog apparatus
hybridized by adding a digital voltmeter? The possibilities are far too numerous.
The present treatment does not even contemplate giving a complete account of
analog computing machines themselves, let alone the combination they may form
with other machines. A large and growing library of good books cover these areas
quite completely. Many of these are written by officials of the Simulation Councils,
who typically have the sort of university connections which should give them
appropriately unbiased viewpoints: viewpoints which a mere company man can
only envy. Perhaps, however, an example or two might be appended here which
will amuse and even edify.
At a large Eastern university, under the guidance of a well-known and gifted
computationalist, a successful project has been reported on whereby the scaling for
an analog installation is done entirely by rote on a digital machine. No guessing or
trial runs at all are involved. Straight from the equations, the digital solution dic-
tates the analog settings which will bring the maximum excursion of every variable
analog voltage to within 20% of the limiting value, Local wags thus proclaim the
discovery at last of a practical contribution by the digital apparatus. Seriously, they
enjoy the ability to “get at” the solutions of the analog during operation.
Some analog men, perhaps over-fond and defensive as regards continuous func-
tions, realty believe that analog operations are generalizations of digital ones, or
that conversely digital operations are special cases of analog ones. What can be done
with such people? They depreciate the importance of the fact that discrete measure-
scales approach continuity in the limit, alleging that infinite processes are already
tacit and available, without passing to the limit, in an analog variable. Pointing for
example to analog selector circuits which can pick out and transmit whichever of a
set of variables is algebraically the greatest of the least, they cite this capability as
broader than the logical sum or the logical product, amounting in fact to infinitely-
many.-valued logic. Selectors followed, for example, by bounding operations serve
directly in the rudimentary case of two-valued logic. On the basis of such reasoning
it is surprising, the argument runs, that analog apparatus is not permitted to make
decisions for itself. It is hard to answer these arguments, especially when dealing with
confinned analog partisans. When cornered on some point of superior digital accom-
plishment, they simply claim the whole digital province as part of their analogs.
Predictions are scheduled for the Tomorrow part of this article, but one such
properly belongs hcrc. While it is agreed that analog and digital techniques will
increasingly cross-fertilize and intcr-relate, it is predicted that the controversy
beween their camps will rage on, good natured but unabated, for years to comc in
spite of hybrid attachments. The serious issue of reliabili1y has recently arisen as

11