| Digital-to-analog converters (DACs) suffer from static and dynamic nonlinearity problems, which degrade their accuracy and performance. Mismatch errors in the analog components restrict the maximum achievable linearity.; This thesis presents various techniques for correcting these errors. It describes a correction process for the nonlinear behavior of DACs, on three different levels: architectural design, circuit design, and layout design.; The main results achieved are listed below: (1) Novel topologies using stochastic approaches to linearize multibit converters are presented. (2) A new method is introduced for avoiding the use of multibit DACs in the main loop of multipath DS analog-to-digital converters (ADCs), which, combined with a novel noise leakage compensation technique, allows the use of low quality inner DACs. (3) A novel correction algorithm is proposed, which is based on the acquisition of the individual DAC errors by means of correlation procedures. The extracted values are used for correction purposes. The technique is capable of background operation. (4) Different circuits are proposed to improve the performance of current-steering DACs. Also, novel layout techniques are shown for reducing the spatial variations of the unit sources. Some of the presented techniques were combined in a prototype chip, designed and fabricated in a 0.35mum CMOS process. Simulation and preliminary measurement results show that they are effective. |
