High Frequency Continuous Time

There is an ever increasing trend towards putting entire systems
on a single chip. This means that analog circuits will have to
coexist on the same substrate along with massive digital systems.
Since technologies are optimized with these digital systems in mind,
designers will have to make do with standard CMOS processes in the
years to come. We address analog filter design from this perspective.
Filters form important blocks in applications ranging from computer
disc-drive chips to radio transceivers.
In this book, we develop the theory and techniques necessary
for the implementation of high frequency (hundreds of megahertz)
programmable continuous time filters in standard CMOS processes.
Since high density poly-poly capacitors are not available in these
technologies, alternative capacitor structures have to be found. Metalmetal
capacitors have low specific capacitance. An alternative is to
use the (inherently nonlinear) capacitance formed by MOSFET gates.
In Chapter 2, we focus on the use of MOS capacitors as integrating
elements. A physics-based model which predicts distortion accurately
is presented for a two-terminal MOS structure in accumulation.
Distortion in these capacitors as a function of signal swing and bias
voltage is computed.
Chapter 3 reviews continuous-time filter architectures in the light
of bias-dependent integrating capacitors. We also discuss the merits
and demerits of various CMOS transconductance elements. The
problems encountered in designing high frequency programmable
filters are discussed in detail.
Chapter 4 considers time-scaling in electrical networks and introduces
a technique called constant-capacitance scaling. We show that
wide-range programmable filters implemented using this technique
are optimal with respect to noise and dynamic range.
Chapter 5 documents the detailed design and layout of a 60 –
350 MHz Butterworth filter implemented in a 0.25 μm digital CMOS
process. A simple circuit arrangement is proposed to keep the
filter bandwidth constant with respect to temperature and process
variations. Simulation results for the filter test chip are shown.
For filters in the frequency range of interest to us, special care has
to be taken to measure the filter response accurately. In Chapter 6,
we present measurement techniques and the implementation results
of the prototype filter chip. This experimental data demonstrates the
effectiveness of the techniques proposed in Chapter 4.
In Chapter 7, we discuss the application of scaling techniques to
other filter architectures.
For very high-quality filters, the simple tuning technique proposed
in Chapter 5 may not guarantee sufficient precision. In Chapter 8, a
more accurate and complex method based on the behavior of a filtercomparator
oscillator is proposed. This scheme converts a filter into
an oscillator. The amplitude and frequency of oscillation are measures
of the center frequency and quality factor of a biquadratic filter
section. Analytical relations for the behavior of the filter–comparator
system are developed and verified through a breadboard prototype.
This book is based on a doctoral thesis (submitted by one of the
authors [S.P.] to Columbia University). It describes the results of
a research project carried out to determine if it is at all possible
to realize robust, production quality VHF filters in standard deep
submicron CMOS technologies. The filter design discussed in this
book was implemented at Texas Instruments (New Jersey).
ACKNOWLEDGMENTS: We are grateful to the staff of Texas
Instruments at Warren for their support - T.R. Viswanathan and K.
Nagaraj have been sources of great encouragement all along. We
thank the students of the Columbia Integrated Systems Laboratory
- especially M. Tarsia, G. Palaskas, N. Krishnapura and A. Dec for
useful discussions. Finally, one of the authors [S.P.] would like to
thank Professors Anthony Reddy and John Khoury - their superb
classes on filter theory and circuit design have been profoundly
influential over the course of creating this book.