A JSSC Classic Paper:
Switched Capacitor Filters Grew
the
Communications Industry
Filtering circuits, the most ubiquitous of electronic signal-processing
functions, made for spirited discussions as analog-to-digital
technology solutions unfolded. Bedrich Hosticka's 1978 paper on
a MOS switched-capacitor filter marked the beginning of the dominance
of switched-capacitor techniques for integrated audio-band filtering.
In the ISSCC 40th anniversary Commemorative Supplement in 1993,
Robert W. Brodersen remembered that in the year following Hosticka's
paper two papers at ISSCC 1979 demonstrated the manufacturability
of the approach in high-volume applications. The circuit served
as part of an icon of the era of IC telecom, when a chip plot
of 30 poles of switched-capacitor filtering by B.J. White, G.
M. Jacobs, and G.F.Landsburg hung in the New York Museum of Modern
Art for a number of years.
Published in the December 1977 JSSC, "MOS-sampled data recursive
filters using switched-capacitor integrators" by Bedrich
Hosticka, Robert Brodersen, and Paul Gray, University of California,
Berkeley, California, is the fourth most frequently cited article
ever published in the JSSC. It has been cited more than 150 times
according the Journal Citation Report 2000 Science Edition. The
concepts were presented first in February of the same year at
ISSCC.
Unknown to Hosticka and his coauthors in 1978, James Clerk Maxwell
had explained the equivalence of a periodically switched capacitor
to a resistor in A Treatise on Electricity and Magnetism, Oxford,
The Clarendon Press, pp. 420-421, 1892. The opportunity for practical
realization of filters came only with the development a zero-offset
electronic switch and an infinite input impedance electronic amplifier.
The MOS transistor was the first device to meet both requirements,
80 years later. Professor Yannis Tsividis at Columbia found this
early prior art.
Prof. David Hodges of UC Berkeley, a co-recipient of the 1983
IEEE Morris Liebmann Technical Field Award for switched capacitor
filters, notes: "SC filters found effective use only below
1 MHz analog. In that range, due to the progress in digital technology,
digital filters used with delta-sigma A/D converters have almost
entirely superseded SC filters for high-order functions.
Switched-C filters still are used for simple filter functions
that can't justify the overhead associated with DSP filters, e.g.
some toys. L-C filters still are used for very simple filters,
such as the signal splitter that separates the DSL channel from
the voice/fax channel on a home telephone line - excellent surge
immunity, zero power consumption!"
As so frequently happens in the industry, a new technology emerged,
had major impact, and then was superseded by subsequent advances
and requirements.
The original paper is available on the Solid-State Circuits Digital
Archive 2002 DVD. See page 13. (IEEE product # JD3755B)
|
The Analog MOS Filter Solution
The paper was a result of my PhD thesis research from 1974 to 1977. When
I started my work at the Department of Electrical Engineering and Computer
Science of the University of California, Berkeley, nobody believed that
you could realize analog integrated circuits using MOS technologies (actually,
all analog integrated circuits realized at that time were using bipolar
technologies). It was the Department of EECS in Berkeley that pioneered
the analog MOS (and later CMOS) design. This work was carried out by a
group guided by Berkeley professors Paul R. Gray, R. W. Brodersen, and
David A. Hodges. They created an environment that attracted gifted young
engineers from all over the world who generated a number of new ideas.
It was this work that laid down the foundations of analog CMOS design.
I was among the first PhD students who joined the group. My work, inspired
mainly by Paul Gray and R.W. Brodersen, introduced a new analog circuit
design technique, which we called "switched-capacitor"technique.
Based on novel analog switched-capacitor filters, we showed that analog
switched-capacitor circuits not only "work"but also deliver
excellent performance. The summary of the work appeared in the December
1977 issue of the IEEE Journal of Solid-State Circuits. I believe that
this paper helped start a revolution in integrated circuit design, the
consequences of which we feel to this very day.
I was glad to be present at the birth of an invention-and to contribute
a little bit. I learned a lot from my professors (to whom I owe very much)
and I have gained many friends as well-my thanks goes to all of them.
Bedrich J. Hosticka
Fraunhofer Institute of Microelectronic Circuits and Systems
Duisburg, Germany
bedrich.hosticka@ims.fhg.de
|