TABLE OF CONTENTS This chapter presents techniques which are commonly used to compensate or hide imperfections of analog circuits. Some of them, like chopper modulation, use mostly analog circuitry to remove a disturbing effect. Others, like successive approximations, extensively use digital correction algorithms to trim analog components or circuits. First, the mostly used techniques are presented. Then, their performances are examined and a classification is made.
The design of analog circuits is rendered difficult by the imperfections imparted by the manufacturing process to the component values. Physical parameters (e.g. oxide thickness, physical dimensions, doping profile) are subject to variations due to instabilities of the fabrication technology, and they reflect on component parameters. The best achievable tolerance of individual component values thus depends on the accuracy of the manufacturing process, and cannot be reduced below a minimum level.
Fortunately, analog design rarely relies on the absolute value of single components, but rather on relative values of several components. The relative values can be made arbitrarily close, i.e. with small tolerances, by using appropriate design techniques like matching. Thus, high-precision circuits can be realized even with poor manufacturing processes.
The most common technique for improving the precision of analog blocks is matching. If the layout of pairs/sets of components is performed carefully following the rules presented below, the statistical dispersion of their values can be reduced.
The following rules should be applied for optimum matching of integrated components [1]:
Same structure
Same temperature
Same shape,...
