This thesis proposes a 50M/s fully passive noise-shaping SAR ADC with mismatch error shaping for advanced driver assistance systems (ADAS). To implement ADAS that provides comfort and safety to drivers, an ADC capable of fast and accurate data process...
This thesis proposes a 50M/s fully passive noise-shaping SAR ADC with mismatch error shaping for advanced driver assistance systems (ADAS). To implement ADAS that provides comfort and safety to drivers, an ADC capable of fast and accurate data processing is essential. The noise-shaping SAR ADC offers the advantage of achieving high resolution by shaping noise while maintaining the low power and small area characteristics of conventional SAR ADCs. However, mismatch noise is not shaped, posing a significant issue even in traditional SAR ADCs. To address this, the thesis employs the MES technique to shape the mismatch noise, which can be relatively simply implemented by adding timing compared to other complex methods such as calibration. However, this approach has the drawback of reducing the dynamic range. To mitigate this, cyclic DEM and LSB cancellation MES techniques are applied. Additionally, to achieve low power and small area, a fully passive cascaded-integrator feed- forward (CIFF) architecture is used without the need for separate amplifiers.