A Wide Dynamic Range Analog-to-Digital Converter Using a Band Limited Predictor-Corrector DPCM Algorithm
by fred j. harris

Differential pulse code modulation (DPCM) is classically used in communication systems to reduce the number of bits required to represent a given input signal. By a simple change in perspective the same process can be used to represent an input signal with an increased number of bits. This perspective likens the predictor loop of the DPCM to the first pass quantizer of a dual rank A-to-D converter and the correlator loop of the DPCM to the second pass of the dual rank converter.
[download document]


On The Performance Of Alternate Architectures For Noise Shaping Modulators Of The Sigma-Delta Type
by fred j. harris, Bill McKnight

This paper presents and compares the performance of alternate architectures for obtaining spectral noise shaping in oversampled ADCs and DACs. These architectures differ from traditional approaches in a number of ways. Standard sigma-delta modulators use multiple zeros at DC to shape the noise spectrum, while these architectures use a structure which control the noise spectrum with zeros distributed over the unit circle spanning the signal bandwidth. The approximately uniform distribution of zeros over the signal bandwidth permits the resampling DAC to operate at lower oversampling rates. We also present and compare relative stability of multiple pole feedback loops using one-bit quantizers.
[download document]


Multirate Digital Signal Processing Techniques to Reduce the Cost of the Analog Signal Conditioning Filters Used at the Boundary Between the Continuous and Sampled Data Representation of Signals
by fred j. harris

Digital filters are used in many standard applications in communication and multimedia systems. These applications include low-pass filtering for noise suppression, band-pass filtering for signal separation, and high-pass filtering for D.C. rejection. Other traditional applications include spectral shaping to control bandwidth and signal wave shape in modulators and demodulators, differentiator filters for FM demodulation, Hilbert Transform filters for single sideband modulation, and pre-emphasis and de-emphasis filters used in FM modulation.
[download document]


Oversampled Digital to Analog Converters With Very Small Computational Load
by fred j. harris, Bill McKnight

We present a non-traditional method of performing the decimation or interpolation required in an oversampled ADC or DAC. This technique uses a very efficient polyphase all-pass structure to perform the resampling task via a sequence of half-band filters. The arithmetic workload is reduced by a factor of five relative to the standard form of interpolation.
[download document]


Improved Performance of Multi-Bit Delta-Sigma Analog to Digital Converters Via Requantization
by fred j. harris, Eric Brooking, Bill McKnight

System performance of an oversampled analog to digital converter (ADC) with feedback noise shaping is limited by the precision of the digital to analog converter ~AC) in the feedback path as well as by the fluter of integrators in the loop(s). Standard designs avoid the DAC precision problem by restricting the ADC and DAC to a single bit while stability and matching considerations limit systems to three loops. This limit in turn defines the oversample ratio for a given effective bandwidth and noise performance. We present a simple modification to the oversampled ADO which avoids these limitations via requantization in the feedback path of the original delta sigma loop structure. This modification results in greater dynamic range than is available from standard configurations.
[download document]


New Architectures With Distributed Zeros For Improved Noise Shaping of Delta-Sigma Analog to Digital Converters
by fred j. harris, Bill McKnight, Bob Caulfield

This paper presents and compares the structure and performance of recent alternate architectures developed for spectral noise shaping in over- sampled analog-to-digital and digital-to-analog converters.These architectures differ from conventional design approaches in two ways. The primary difference is in the distribution of transmission zeros in the noise transfer function, while the secondary difference is the manner in which the architecture forms these transmission zeros.
[download document]