Degree Type


Date of Award


Degree Name

Doctor of Philosophy


Electrical and Computer Engineering

First Advisor

Randall L. Geiger


Digitally-based signal generation systems are attractive for their ability to quickly change phase, frequency and amplitude with small settling transients. The stability and variety of the signals produced is also greater than those of analog designs. The limiting factor to the spectral quality of digital systems is the Digital-to-Analog Converter (DAC) which provides the continuous-time output. A great deal of work is now underway to improve the frequency-domain performance of these DACs;The major contributor to distortion in high-speed DACs tailored for wave-form generation is the lack of linear transitions in the output signal. These dynamic errors have numerous circuit-level origins, and are only vaguely understood. Most of the work done in this area to date has centered on circuit-level cures to individual error sources, and has ignored system-level solutions;The work reported here explores the origins of spectral distortion in high-speed DACs, with emphasis on transition errors. In contrast to previous literature, estimates of the distortion levels incurred by dynamic errors are made. Existing cures are then analyzed, and system-level techniques are investigated in an effort to overcome a wider variety of nonlinearities;The outcome of this work is a novel closed-loop conversion method which measures and adjusts for a wider class of DAC and system nonidealities than previous designs. This is accomplished by evaluating the spectrum of the signal being generated and "predistorting" the DAC input to compensate for errors. This system was constructed from commercial components and demonstrates the algorithm's potential by substantially increasing the dynamic range of an off-the-shelf DAC. The achieved performance is as good or better than currently available systems. The viability of fully-integrated designs is discussed, as are applications in production testers and integrated test systems.



Digital Repository @ Iowa State University,

Copyright Owner

George R. Spalding, Jr.



Proquest ID


File Format


File Size

170 pages