L. Toth

General results for resistive noise in active RC and MOSFET-C filters

Resistive noise in active RC and MOSFET-C filters is examined. The effects of impedance and frequency scaling are shown. The relationship between sensitivity and noise is considered. A bound for the product of resistive power dissipation and noise power spectral density is derived. Topology-independent and transfer function-independent bounds for the dynamic range, in filters with power dissipation constraint, are developed.

Noise analysis of ideal switched-capacitor networks

This paper presents a method for computing the noise power spectral density (PSD) in switched-capacitor networks, Explicit formulas are developed for both white and 1/f noise sources. The 1/f noise is handled directly in the formulation without having to approximate its PSD by using a noise shaping filter. The technique eliminates the need for computing eigenvalues or matrix exponentials as well as solving bilateral matrix equations, and thereby leads to a numerically efficient algorithm. The noise sources considered here are the ON-resistances of MOS switches and the input referred noise PSDs of operational amplifiers (opamps). Experiments are carried out to demonstrate the validity of the formulation for both white and 1/f input noise.

Generalization of the principle of chopper stabilization

The concept of traditional chopper-stabilized amplifiers is extended to general linear dynamical and certain types of nonlinear circuits. Simulation results are given to confirm the theory.

Noise analysis of externally linear systems

Noise analysis in externally linear time-invariant systems is complicated by the facts that noise sources may not be stationary and the relationships between the noise sources and the output are generally nonlinear. In this paper, we present a general noise analysis method which can be used to speed up repetitive noise calculations for a given system, and, most importantly, for theoretical investigations of its noise behavior, possibly leading to low-noise design techniques. A current mode syllabic companding and two log-domain filters are analyzed as illustrative examples.

Generalized chopper stabilization

The concept of traditional chopper-stabilized amplifiers is extended to linear dynamical and certain types of nonlinear circuits. Simulation results are given to confirm the theory.

Noninvasive techniques for syllabic companding in signal processors

Two novel methods are described for syllabic companding in signal processors which make state variable updating unnecessary, for piecewise constant input and output gains. Measurements and simulations are used as confirmation of the validity of the proposed techniques.

Analytical approach for the exact phase noise analysis of oscillators

The output noise power spectral density of a class of oscillators is calculated analytically and exactly. An example is shown as an illustration. The analytical results derived here agree with measurements and simulations published elsewhere.

Two "noninvasive" techniques for syllabic companding in signal processors

Two novel methods are described for syllabic companding in signal processors. These methods make state variable updating unnecessary, for piecewise constant input and output gains. Measurements and simulations are used as confirmation of the validity of the proposed techniques.

Noise in Externally Linear Filters

This paper presents a systematic noise analysis method for externally linear filters, suitable for hand calculations. As an illustrative example, a log-domain filter is analyzed. Applying this method to an abstract topology, bounds for signal-to-noise ratio are derived for a class of instantaneous companding filters, when total capacitance or total power are limited.

A new approach for the z-domain analysis of linear periodically time-varying discrete time systems

A report is presented on a closed matrix formalism for analyzing linear periodically time-varying (LPTV) systems specific by either the input-output equation or the canonical realization. It is shown that the two structures usually behave differently if the coefficient sequences are the same, but for a special type of LPTV system the same coefficients are suitable for both the direct-form and canonical realizations. In this case the transfer characteristics can be expressed analytically instead of using a matrix description. A switched capacitor example illustrates how easy it is to analyze interconnections of LPTV systems on the basis of the input-output equations of the building blocks by applying the method presented.<<ETX>>