[EAS]Analog and Digital

[pjk]

Subject:   Analog and Digital

(from INNOVATION, 5 February 2003)

DIGITAL'S OUT, ANALOG'S IN
"Digital" has long been a code-word for "cool" and "cutting edge,"
but  National Semiconductor CEO Brian Halla predicts a renaissance
for old-fashioned analog chips, triggered by the rise in electronic
monitoring of all types and the increased use of the Internet for
transferring images and graphics. "The only things on the face of
the planet that use zeros and ones are microprocessors and digital
signal processors," says Halla. "It's  fine to do zeros and ones for
spreadsheets and that's why the PC uses the least amount of analog.
But we're not doing spreadsheets anymore. We're  doing digital
photography. We're downloading images and graphics from the 
Internet, and we're doing more and more stuff wirelessly. All of
that is  analog." Along with returning to its analog roots, National
Semi is working  on its own cutting edge technologies. Halla
envisions a smart card with biometric information, which instead of
using a digital image of a thumbprint (think ones and zeros) to
activate it, employs a special analog  technology that could
actually prove more secure: "(I)f you ever lose it, (the card)
doesn't have your thumbprint so it's useless. We're working with a
particular technology where you don't leave your thumbprint; you rub
it."  (CNet News.com 29 Jan 2003)
http://news.com.com/2008-1082-982481.html

---------------------------------------------------------------------
As an "analog guy" I always perk up when I read something like this,
but this summary doesn't make much sense. Not the fault of the
INNOVATION editors, the CNET.com story isn't very clear either, but
does have some pertinent observations. Analog design just isn't as
clearly separate as it used to be, but is part of sensors,
actuators, of analog-to-digital converters, of a lot of enabling
circuitry that allows digital chips to perform meaningful functions
in an analog world. And as Halla points out, analog chips are small
and have a high profit margin. 
As generalist problem solvers, analog designers tend to be
individualists. With new sensing tasks arising from biomedical
engineering, and new non-silicon device technologies spawned by
molecular electronics research, it will likely be analog designers
that recognize their new "rules of design" and, like Bob Widlar in
his day, grow them into larger operating entities.  --PJK

----------------------------------------------------------------
P.S.: If you haven't heard of Bob Widlar, back in April of 1991,
when I first heard of his death, I wrote this on the spur of the
moment for my class:

Only just now did I learn from an IEEE news item that Robert J.
Widlar died Feb. 27th at age 53. No cause of death was given.

After graduating from the Univ. of Colorado, Widlar joined Fairchild
and in 1965 designed the uA709, one of the first linear integrated
circuits of commercial significance. It was a turning point in
circuit design and implementation, and is still in the catalog even
today. Prior to the 709 monolithic op amps consisted only of NPN
transistors and resistors (e.g. the early RCA integrated circuits).
The 709 was the first op amp utilizing high-beta lateral PNPs which
solved many dc voltage level shifting problems and provided good
constant-current biasing. Monolithic op amp performance, while not
as good as the more expensive discrete component op amps, was now
adequate for many systems. While at Fairchild he also designed the
uA723 regulator, a real classic, also in the catalog to this day.

In 1966 he joined National Semiconductor, where he designed the
LM301, the first op amp with short circuit protection, freedom from
latchup and a full 30V differential input capability, though it was
still externally compensated. (It was back at Fairchild that others
designed the 741 in response to the 301.) At National he contributed
many designs, including the LM308 op amp in 1969, with a 2nA input
bias current, the first use of super-beta transistors in op amp
input stages. Super-beta behavior had been observed earlier, but it
took Widlar's special circuit configuration to cope with the low
collector voltage requirements of these devices in a practical
circuit.

Widlar's ingenuity brought a whole new approach to designing on
silicon, where transistors are cheap but other components are
expensive, where absolute parameters are hard to control, but
matching is easy. He thought in terms of patterns, didn't "get lost
in the equations by the third transistor" and was generally
disdainful of computer aids. "... But you can't mechanize the
(linear) process because once you do that, you stop inventing. And
then you very much restrict what you're going to do" he was once
quoted. Whereas a lot of earlier linear ICs looked like discrete
component circuits on a chip, his looked distinctively different. He
was the first who really understood the linear IC medium.

After retiring from National Semi in his thirties to live in Mexico,
he got restless and resumed some work for National, including the
LM10 op amp with voltage reference, which broke new ground in low
voltage design with operation down to a total supply voltage of 1.0
volt. When using it for a regulator, you can power the whole IC
from the base-emitter drop of a Darlington pass transistor, ideal
for a high-voltage floating regulator. And consider that the LM10
works down to -55 deg.C, where a base-emitter drop is approaching
one volt! He joined Linear Technology in 1981, and more recently
consulted again for National Semi.

IC engineering is not exactly a breeding ground for "folk heroes"
and "originals", but Bob Widlar was both. I am sad that he is no
longer working amongst us.

  --pjk