IEEE Solid-State Circuits Magazine - Winter 2016 - 24

filters and was manufactured in the
same NMOS process as the E13. It
went into production in 1981. The
transmit filter contained a biquad
for the 50/60-Hz notch and a fifthorder filter (two biquads and a firstorder filter) for the anti-aliasing
section. The receive filter had a similar fifth-order filter.
Care was also taken in the design
of the amplifiers and capacitors to
ensure reduced idle channel noise.
As a result, the total idle channel
noise was greatly reduced and the
design was much more tolerant to
process variations. Performance
results for the E99 versus E13 +
access chip are shown in Table 1.
Unlike other commercially available filter codecs at that time, the
Nortel E99 design incorporated many
other functions of the line card. As a
result, this chip was no longer a component of the line cardbut integrated
a large part of the line card on chip.
This greatly improved the performance of the line card and reduced
its cost as well. Subsequently, a thirdgeneration CMOS version (code X11)
of the "filter codec" chip followed,
integrating even more functionality of the line card onto the chip.
Eventually, a "world line card" IC
was developed by the late 1980s and
manufactured at Nortel.
The end result of this innovation is that more than 100 million
of these chips were produced and
installed in DMS-100 switching systems. This allowed Nortel to become
the world leader in digital communications and basically kicked-off the
digital revolution in voice communications. Switched-capacitor filters
became the world's first choice for
implementing analog filters in silicon. In 1978 Nortel had a big problem
of critical importance, and Dr. Copeland's fundamental solution enabled
the company to build and sell tens
of billions of dollars worth of DMS100 digital switches. Sure, during

W I N T E R 2 0 16

Nortel had an extensive research program
on CCDs at the time, and Dr. Copeland,
Prof. Chong Chan, and their graduate students
at Carleton University were involved.
the development of DMS-100 there
were other holes that needed to be
plugged, but this one would have
been a "show stopper," and Dr. Copeland's solution put a cork in it.
While working with Dr. Copeland,
I observed that he has the ability to
see problems and solutions with
such clarity. He would often say,
"Well intuitively we can see 'such
and such'," and then take a connective jump to the next step in our
work. It might have been intuitive to
him, but the rest of us needed time
to catch up and master new developments! What impressed me most
was the scope of his thinking when
trying to understanding or solve
a problem. It was never about trying to build a better mousetrap for
Miles, instead it was always about
problem solving at a more fundamental level and understanding the

TAblE 1. E99 vErsus E13
PErformANcE.

PArAmETErs

E99
PErformANcE
vErsus E13 +
AccEss chiP

Idle channel
noise

6 dB lower

Harmonic
distortion

6 dB lower

Low-level gain
tracking

2× better

Absolute gain

4× better

Overall chip yield

25% better

Chip size

20% reduction

Line card
reliability

3× better

IEEE SOLID-STATE CIRCUITS MAGAZINE

impact and opportunities presented
by the solution to such problems.
Prof. Copeland once told me that
he thought of the switched capacitor
idea while drinking coffee on Sunday
morning while still in his pajamas.
Perhaps this gives you a better idea
of what he could do on a work day.

References

[1] J. T. Caves, M. A. Copeland, C. F. Rahim,
and S. D. Rosenbaum, "Sampled analog
filtering using switched capacitors as
resistor equivalents," IEEE J. Solid-State
Circuits, vol. SC-12, no. 6, pp. 592-599,
Dec. 1977.
[2] B. J. Hosticka, R. W. Brodersen, and P. R,
Gray, "MOS sampled data recursive filters
using switched capacitor integrators,"
IEEE J. Solid-State Circuits, vol. SC-12, no.
6, pp. 600-608, Dec. 1977.

About the Author
Terry Caves (terry_caves@hotmail.
com) has spent more than 35 years
in the high-tech industry in Canada
and China. He graduated from Algonquin College in Ottawa, Canada, and
joined the semiconductor integrated
design department at Northern Telecom/Bell-Northern Research in 1969.
He was involved in the early development work on charge-coupled imaging and serial memory devices. He
was the lead engineer on the world's
first integrated voice filter and codec
(E13). He held a variety of other positions at Nortel, including director of
operations for Nortel's Semiconductor Fab in Ottawa and president of
Shanghai/Nortel, an IC design joint
venture in China. His career also includes a venture into photonics as
director of engineering and manufacturing at JDS-Uniphase, and vice
president of engineering and manufacturing at Zenastra Photonics.

Table of Contents for the Digital Edition of IEEE Solid-State Circuits Magazine - Winter 2016