Accurate self characterization and correction of A/D converter performance

A new algorithm based on a code density test to measure Integral Non-linearity (INL) of Analog-to-Digital (A/D) converters is introduced. This algorithm uses as an input two related but unknown 'ramp-like' signals that can be practically generated on-chip and the corresponding code-density histogram. The algorithm inherently characterizes both the A/D converter and the inconsequential input. This algorithm finds applications in self-calibrating A/D converters as well as providing capability for Built-In Self-Test, where generation of an on-chip ramp input test signal of 8bits accuracy or more is a challenge.