Parameterization of Nonlinearity for Efficient Estimation in ADC Testing

In Maximum Likelihood (ML) estimation of ADC parameters, the estimation of static transfer characteristic has key importance. However, the description of static transfer characteristic demands to handle numerous values. Without parameterization, the non-linearity of an N-bit converter can be described using 2(N) - 1 integral nonlinearity (INL) values. Nevertheless for real ADCs the INL values are not independent, the information regarding the nonlinearity can be compressed. This paper enumerates multiple methods to measure and approximate the static transfer characteristic of the ADCs, and evaluates their efficiency. The results are expressed paying attention to the number of parameters and using standard measures for the approximation error (e.g. the l(infinity) norm of the error vector).