TL;DR: It is shown that the arbitrary shaping of mismatch noise of a segmented DAC is possible by dynamic element matching techniques requiring only relatively simple digital element selection hardware.
Abstract: It is shown that the arbitrary shaping of mismatch noise of a segmented DAC is possible by dynamic element matching techniques requiring only relatively simple digital element selection hardware. High order noise shaping promises converters with increased bandwidth and resolution. Examples of first and second order lowpass and bandpass functions are given.
TL;DR: A 13-bit 80-kHz baseband analog-to-digital converter suitable for use in applications such as the ISDN U-interface is described, with two-stage third-order noise shaping.
Abstract: A 13-bit 80-kHz baseband analog-to-digital converter suitable for use in applications such as the ISDN U-interface is described. Two-stage third-order noise shaping permits the use of asampling frequency of only 2.56 MHz. The circuit has been implemented using conventional single-ended switched-capacitor techniques in a 1.5- mu m CMOS process. >
TL;DR: A stereo D/A converter for digital audio applications is presented which obtains 16 bit resolution from a one bit converter, using a code conversion technique based upon oversampling and noise shaping.
Abstract: A stereo D/A converter for digital audio applications is presented which obtains 16 bit resolution from a one bit converter, using a code conversion technique based upon oversampling and noise shaping. Linear-phase low-pass filtering is implemented in the code converter to allow simple analog output filtering.
TL;DR: A new architecture applicable to multibit noise-shaping D/A conversion that relies on the principle that only a few LSBs of an oversampled PCM digital signal need to be noise shaped in order to attain substantially the same performance as with an ordinary single stage or multistage architecture.
Abstract: In this paper, we describe a new architecture applicable to multibit noise-shaping D/A conversion. Significant digital hardware savings are attained due to the inherent reduced bit width required for the data path in the noise shaper. In addition, the error cancellation is inherently accurate, since it is performed entirely in the digital domain. The architecture relies on the principle that only a few LSBs of an oversampled PCM digital signal need to be noise shaped in order to attain substantially the same performance as with an ordinary single stage or multistage architecture.