Oversampled analog-to-digital (A/D) converter architectures have been receiving increased attention for high-precision A/D converters. These architectures offer the means of exchanging resolution in time for that in amplitude. Among these oversampled A/D converters, delta-sigma modulators are the most popular method used due to their simplicity in the analog circuitry. The analog integrators in delta-sigma modulators suffer from non-idealities such as capacitor mismatches and finite op-amp gain. In the dual quantizer A/D converters, the system relies on the perfect matching of the analog and digital transfer functions to cancel the quantization noise. However, the non-ideality of the analog parameters makes this matching hard to achieve. In this thesis, an off-line adaptive scheme is presented to estimate the non-ideal parameters of the analog section for the second-order delta-sigma modulator. These estimates are then used in the digital part to reduce the quantization noise. The least-mean- square (LMS) algorithm is used to adaptively estimate the analog parameters. / Graduation date: 1993