Digital LMS adaptation of analog filters without gradient information

The least mean square (LMS) algorithm has practical problems in the analog domain mainly due to DC offset effects. If digital LMS adaptation is used, a digitizer (analog-to-digital converter or comparator) is required for each gradient signal as well as the filter output. Furthermore, in some cases the state signals are not available anywhere in the analog signal path necessitating additional analog filters. Here, techniques for digitally estimating the gradient signals required for the LMS adaptation of analog filters are described. The techniques are free from DC offset effects and do not require access to the filter's internal state signals. Digitizers are required only on the input and error signal. The convergence rate and misadjustment are identical to traditional LMS adaptation, but an additional matrix multiplication is required for each iteration. Hence, analog circuit complexity is reduced but digital circuit complexity is increased with no change in overall performance making it an attractive option for mixed-signal integrated systems in digital CMOS. Signed and subsampled variations of the adaptive algorithm can provide a further reduction in analog and digital circuit complexity, but with a slower convergence rate. Theoretical analyses, behavioral simulations, and experimental results from an integrated filter are all presented.

[1]  R. Castello,et al.  An eighth-order CMOS lowpass filter with 30-120 Mhz tuning range and programmable boost , 2001, Proceedings of the 26th European Solid-State Circuits Conference.

[2]  K. Nagaraj,et al.  A high speed, low power PRML read channel device , 1995 .

[3]  Anthony Chan Carusone,et al.  Analogue adaptive filters: past and present , 2000 .

[4]  Stephen H. Lewis,et al.  A CMOS adaptive continuous-time forward equalizer, LPF, and RAM-DFE for magnetic recording , 1999 .

[5]  A. Peterson,et al.  Transform domain LMS algorithm , 1983 .

[6]  Anthony Chan Carusone,et al.  A 5th order Gm-C filter in 0.25 /spl mu/m CMOS with digitally programmable poles and zeroes , 2002, 2002 IEEE International Symposium on Circuits and Systems. Proceedings (Cat. No.02CH37353).

[7]  G. Feygin,et al.  A 160 MHz analog equalizer for magnetic disk read channels , 1997, 1997 IEEE International Solids-State Circuits Conference. Digest of Technical Papers.

[8]  J. Hauptmann,et al.  An integrated adaptive analog balancing hybrid for use in (A)DSL modems , 1999, 1999 IEEE International Solid-State Circuits Conference. Digest of Technical Papers. ISSCC. First Edition (Cat. No.99CH36278).

[9]  S. Haykin,et al.  Adaptive Filter Theory , 1986 .

[10]  F. J. Kub,et al.  Analog CMOS implementation of high frequency least-mean square error learning circuit , 1995 .

[11]  Audra E. Kosh,et al.  Linear Algebra and its Applications , 1992 .

[12]  G. Moschytz,et al.  Adaptive switched capacitor filters based on the LMS algorithm , 1990, IEEE International Symposium on Circuits and Systems.

[13]  David A. Johns,et al.  Median-based offset cancellation circuit technique , 1992, [Proceedings] 1992 IEEE International Symposium on Circuits and Systems.

[14]  David A. Johns,et al.  Automatic tuning of continuous-time integrated filters using an adaptive filter technique , 1991 .

[15]  D. Hirsch,et al.  A Simple Adaptive Equalizer for Efficient Data Transmission , 1970 .

[16]  W. Martin Snelgrove,et al.  Orthonormal ladder filters , 1989 .

[17]  Takashi Morie,et al.  A 200-MHz seventh-order equiripple continuous-time filter by design of nonlinearity suppression in 0.25-/spl mu/m CMOS process , 2002 .

[18]  Sang-Soo Lee,et al.  A BiCMOS continuous-time filter for video signal processing applications , 1998 .

[19]  David A. Johns,et al.  Comparison of DC offset effects in four LMS adaptive algorithms , 1995 .

[20]  J.P. Keane,et al.  An adaptive analog noise-predictive decision-feedback equalizer , 2000, 2000 Symposium on VLSI Circuits. Digest of Technical Papers (Cat. No.00CH37103).

[21]  Michael T. Orchard,et al.  Adaptive filtering using filter banks , 1996 .

[22]  B. Widrow,et al.  Stationary and nonstationary learning characteristics of the LMS adaptive filter , 1976, Proceedings of the IEEE.

[23]  M. Tarsia,et al.  Design considerations and implementation of a programmable high-frequency continuous-time filter and variable-gain amplifier in submicrometer CMOS , 1999, IEEE J. Solid State Circuits.

[24]  Asad A. Abidi,et al.  A 160-MHz analog front-end IC for EPR-IV PRML magnetic storage read channels , 1996 .

[25]  S. Thomas Alexander,et al.  Adaptive Signal Processing , 1986, Texts and Monographs in Computer Science.

[26]  Anthony Chan Carusone,et al.  Obtaining digital gradient signals for analog adaptive filters , 1999, ISCAS'99. Proceedings of the 1999 IEEE International Symposium on Circuits and Systems VLSI (Cat. No.99CH36349).

[27]  C. Rohrs,et al.  A stability problem in sign-sign adaptive algorithms , 1986, ICASSP '86. IEEE International Conference on Acoustics, Speech, and Signal Processing.

[28]  C.-P. J. Tzeng.  An adaptive offset cancellation technique for adaptive filters , 1990, IEEE Trans. Acoust. Speech Signal Process..

[29]  D. Luenberger.  Optimization by Vector Space Methods , 1968 .

[30]  P. Siniscalchi,et al.  A 450 Mb/s analog front-end for PRML read channels , 1999, 1999 IEEE International Solid-State Circuits Conference. Digest of Technical Papers. ISSCC. First Edition (Cat. No.99CH36278).

[31]  Huang Qiuting.  Offset compensation scheme for analogue LMS adaptive FIR filters , 1992 .

[32]  David A. Johns,et al.  Continuous-time LMS adaptive recursive filters , 1991 .

[33]  Behzad Razavi,et al.  A 125-MHz mixed-signal echo canceller for Gigabit Ethernet on copper wire , 2001 .

[34]  Simon Haykin,et al.  Adaptive filter theory (2nd ed.) , 1991 .