Low Power High Dynamic Range A/D Conversion Channel

This work concerns the analysis of an adaptive analog-to-digital (A/D) conversion channel for use with a micro electromechanical system (MEMS) microphone for audio applications. The adaptive A/D conversion channel uses an automatic gain control (AGC) for adjusting the analog preamplifier gain in the conversion channel in order to avoid distortion for large input signals. In combination with a low resolution A/D converter (ADC) and a digital gain block, the adaptive A/D conversion channel achieves an extended dynamic range beyond that of the ADC. This in turn reduces the current consumption of the conversion channel in comparison to a static A/D conversion channel; this at the cost of a reduced peak signal-to-noise ratio (SNR). The adaptive A/D conversion channel compensates for the change in analog gain by a digital gain, thus achieving a constant channel gain in the full dynamic range. However, this compensation results in the generation of audible transient errors in the conversion channel output. The adaptive conversion channel is modeled in order to analyze the factors that impact the performance of the conversion channel, including the generation of the transient error. To evaluate the audibility of the transient errors, an objective method based on the Perceived Evaluation of Audio Quality (PEAQ) method is investigated and compared with a subjective evaluation. The results of the evaluation provide key knowledge about the transient glitches from both a system and psychoacoustical point-of-view. Based on this knowledge, a new method is proposed for the reduction of the transient glitches, based on linear extrapolation of the channel output signal. The design of a low power continuous-time (CT) Delta-Sigma (∆Σ) ADC for use in the adaptive A/D conversion channel is also presented. When designing a CT ∆Σ ADC, the choice of e.g. integrator topology, feedback waveform, feedback type, noise transfer function, and quantization levels, results in a large design space, both at the modulator and circuit level. A new optimization method is presented, that seeks to minimize the current consumption of the ADC. Based on an analysis of the modulator circuits and loopfilter, the optimization method determines a theoretical minimum current solution based on a set of performance requirements. Furthermore the use of current mode feedback in combination with active-RC integrators in the CT ∆Σ ADC is investigated as a method for reducing the current consumption of the ADC, without sacrificing the noise performance of the ADC. The main scientific contributions described in this thesis can be divided into two parts: contributions related to AGC audio systems, and contributions related to low power CT ∆Σ ADC design. In the area of AGC audio systems, the main contributions are: an overview of the challenges in applying AGC to audio systems; a proposed objective method for evaluating the audibility of the transient glitches generated by the adaptive A/D conversion channel; and method for reducing the transient glitches generated by the adaptive A/D conversion channel. In the area of low power CT ∆Σ ADC design a substantial contribution is given. The presented optimization method and the use of current mode feedback identi- fies the possibilities of achieving a low power design by considering the modulator and circuit design as interdependent rather than two separate parts of the design.