The design of mixed-signal integrated circuits has evolved from simple analog and digital circuits operating on the same silicon substrate to the point that now we have complete system on a chip solutions for communication systems. The levels of integration needed to remain cost effective in today's integrated circuit (IC) market require careful use of all the available die space. The current trend of digital to analog converter (DAC) design has focused on maximizing speed and linearity for high performance telecommunications systems. The circuit design methods used to achieve very high sample rates require the use of large amounts of die space. This thesis presents a 10-bit DAC that has been optimized for area, while still maintaining accurate operation at low frequencies. To achieve 10-bit performance, an ultra high gain op-amp is introduced for various servoing applications in the DAC. The architecture chosen for the DAC will show an optimization of required die size and performance when compared to other architectures. The DAC was fabricated in a standard digital 0.18 μm CMOS process. The DAC occupies 0.0104 mm² (110 μm x 94 μm), and only consumes 2.8 mW of power. In addition to the 10-bit DAC, a design is presented for a 13-bit DAC which occupies 0.020 mm², and requires only the addition of a minimum number of devices to the 10-bit DAC. / Graduation date: 2002

Oversampling and noise-shaping methods for digital-to-analog (D/A) conversion have been widely accepted as methods of choice in high performance data conversion applications. In this thesis, the fundamentals of D/A conversion and oversampling D/A conversion were discussed, along with the detailed analysis and comparison of the reported state-of-the-art oversampling D/A converters. Conventional oversampling D/A converters use 1-bit internal D/A conversion. Complex analog filters and/or large oversampling ratios are usually needed in these 1-bit oversampling D/A converters. Using multi-bit internal D/A conversion, the analog filter can be much simpler and the oversampling ratio can be greatly reduced. However, the linearity of the multi-bit D/A converter has to be at least the same as that required by the overall system. The dual-quantization technique developed in the course of this research provides a good alternative for implementing multi-bit oversampling D/A converters. The system uses two internal D/A converters; one is single-bit and the other is multi-bit. The single-bit D/A converter is used in a path called the signal path while the multi-bit D/A converter is used in a path called the correction path. Since the multi-bit D/A converter is not directly placed in the signal path, its nonlinearity error can be noise shaped by an analog differentiator so that the in-band noise contribution from the nonlinearity error is very small at the system output, greatly reducing the linearity requirement on the multi-bit internal D/A converter. An experimental implementation of an oversampling D/A converter using the dual-quantization technique was carried out to verify the concept. Despite about 10 dB higher noise than expected and the high second-order harmonic distortion due to practical problems in the implementation, the implemented system showed that the corrected output had more than 20 dB improvement over the uncorrected output in both signal-to-noise ratio and dynamic range, demonstrating the validity of the concept. / Graduation date: 1996

Oversampled data converters are becoming increasingly popular for high-precision data conversion. There have been many publications on oversampled analog-to-digital (A/D) converters but relatively few on oversampled digital-to-analog (D/A) converters. In this thesis, issues concerning the analysis and design of the oversampled D/A converters are addressed. Simulation tools and analytical methods are discussed. A novel dual-quantization technique for achieving high-precision D/A conversion is proposed. A design example is presented to demonstrate that in many aspects the proposed technique is superior to existing techniques. The thesis is divided into four chapters. Chapter 1 is an introduction to the general concepts of Nyquist-rate and oversampled data converters. Chapter 2 describes some building blocks to be used in oversampled D/A converters and gives both theoretical and simulation methods for analzying them. Chapter 3 describes the proposed dual-quantization D/A converters, including the structure, the associated design issues and an example to verify the validity of this technique. Finally, Chapter 4 summarizes the properties of the simulated system and proposes some future research work. / Graduation date: 1992

No description available.

Thesis (Ph. D.)--Oregon State University, 2006. / Printout. Includes bibliographical references (leaves 86-88). Also available on the World Wide Web.

Thesis (Ph. D.)--Oregon State University, 2007. / Printout. Includes bibliographical references (leaves 89-91). Also available on the World Wide Web.

The goal of this research is to design a low power, high speed, 10-bit, 250 MHz digital-to-analog converter. For high speed application, the DAC is implemented in thermometer-code based segmented DAC. An optimal switching scheme is used in this design. The switching scheme can compensate the gradient error in thermometer-code DAC arrays.This DAC is implemented in a 0.18£gm 1P6M mixed-signal CMOS process provided by TSMC.

No description available.

The design of a high speed circuit accepting a bipolar analog signal with a 3 db bandwidth of 20 MHZ and an eight bit unipolar gain control signal is presented in this thesis. The system produces the product of these two signals at a rate of one digital byte every 25 nsec. At the heart of the system are two multiplying digital to analog converters (DACs) operating in parallel. The circuit design was based on a statistically validated model for a multiplying DAC. This circuit could be used for controlling the intensity of each picture element (i.e. pixel) for many existing video display systems.

Delta-sigma modulators are currently a very popular technique for making high-resolution analog-to-digital converters (ADCs) and digital-to-analog converters (DACs). These oversampled data converters have several advantages over conventional Nyquist-rate converters, including an insensitivity to many analog component imperfections, a simpler antialiasing filter and reduced accuracy requirements in the sample and hold. A recent development in the realm of delta-sigma-based ADC and DAC systems is the use of multilevel (as opposed to binary) quantization. This development owes its existence to the discovery of a variety of techniques which cause linearity errors of the embedded multilevel DAC to be attenuated in the frequency band of interest. This thesis presents several methods for shaping the DAC element mismatch error and reducing the dynamic error in the band of interest. To demonstrate the effectiveness of the proposed algorithms, a current-mode unit element DAC is designed and used as a test bed. Both theoretical analysis and experimental results show that these methods can greatly attenuate the noise in the band of interest. The methods presented in this thesis will allow high performance, high-frequency wideband delta-sigma modulators to be constructed. / Graduation date: 1998