A high-accuracy, DC-calibrated, monolithic, delta-sigma analog-to-digital converter.
AuthorEarly, Adrian Bruce.
AdvisorHamilton, Douglas J.
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractDelta-Sigma Analog-to Digital Converters have recently become important for providing high resolution with monotonicity and reasonable signal-to-distortion ratings without the need for laser trimming techniques. This has come about because of the recent ability to combine both extensive digital computation power, and switched-capacitor analog circuitry on a monolithic chip. Delta-Sigma converters have primarily been used, however, in signal processing applications, notably digital audio, but not for instrumentation. The purpose of this dissertation is to provide a high accuracy, DC-accurate, Delta-Sigma Analog-to-Digital converter in monolithic form. Autocalibration gives endpoint correction, and chopper stabilization minimizes the effect of parameter shifts, drift, and flicker noise. A digital filter, needed for all Delta-Sigma converters, serves as a signal processor to reject out-of-band noise and resonant responses of the external system. A 3-micron, double-poly CMOS process is used. Power requirements are +/- 5 Volts. A six-pole Gaussian IIR digital filter is chosen for good transient response and no overshoot. The filter algorithm and hardware solve the difference equations of a low-pass switched-capacitor prototype filter in digital form. Due to the low bandwidth needed, an area-efficient shift-and-add architecture is used. The area is further reduced with a novel multiplication algorithm, and the logic is reused to perform the calculations required for calibration. The system level device performance is verified in FORTRAN. The analog subcircuits are simulated over process and temperature corners in HSPICE. Measurements show differential and integral linearlity, DC accuracy and noise near the 20-bit level. Power supply rejection, and out-of-band signal attenuation are good, and the step response is monotonic. The circuit is marketed as Crystal Semiconductor CSC5503 and CSC5501 (20 and 16-bit resolutions, respectively).
Degree ProgramElectrical and Computer Engineering