Maximum likelihood estimation of ADC parameters

OData support
Dr. Kollár István
Department of Measurement and Information Systems

Analog-to-digital conversion is essential in embedded systems, where perception of the

physical environment and digital processing of signals is required simultaneously, and

mostly in real-time. Test methods for analog-to-digital converters (ADCs) have been improved

in parallel with the development of the ADC circuits. Several techniques are available

to observe the static and dynamic behavior of the converters. Driving the ADC under

test using a sinusoidal excitation signal, and fitting sine wave to the measurement record is

a very important and meaningful test procedure. This method has been standardized, and

appears in documents released by the IEEE and the IEC. However, it is possible to improve

this technique, using more accurate estimation of the excitation signal. Maximum likelihood

(ML) estimation provides better estimators, thus datasheet quantities of the ADC

under test can be calculated with improved precision. This paper focuses on the practical

realization of ML estimation for ADC testing: examines several challenges concerning the

implementation of this method and proposes solutions for these problems. Investigates the

properties of the ML estimators in comparison with the properties of the standardized least

squares (LS) estimators. Finally, implementation of the ML estimation (and multiple other,

standardized ADC test methods) in two different environments (MATLAB and LabVIEW)

is presented.


Please sign in to download the files of this thesis.