Methodology for testing high-performance data converters using low-accuracy instruments

摘要

There has been explosive growth in the consumer electronics market during the last decade. As the IC industry is shifting from PC-centric to consumer electronics-centric, digital technologies are no longer solving all the problems. Electronic devices integrating mixed-signal, RF and other non-purely digital functions are becoming new challenges to the industry. When digital testing has been studied for long time, testing of analog and mixed-signal circuits is still in its development stage. Existing solutions have two major problems. First, high-performance mixed-signal test equipments are expensive and it is difficult to integrate their functions on chip. Second, it is challenging to improve the test capability of existing methods to keep up with the fast-evolving performance of mixed-signal products demanded on the market. The International Technology Roadmap for Semiconductors identified mixed-signal testing as one of the most daunting system-on-a-chip challenges;My works have been focused on developing new strategies for testing the analog-to-digital converter (ADC) and digital-to-analog converter (DAC). Different from conventional methods that require test instruments to have better performance than the device under test, our algorithms allow the use of medium and low-accuracy instruments in testing. Therefore, we can provide practical and accurate test solutions for high-performance data converters. Meanwhile, the test cost is dramatically reduced because of the low price of such test instruments. These algorithms have the potential for built-in self-test and can be generalized to other mixed-signal circuitries. When incorporated with self-calibration, these algorithms can enable new design techniques for mixed-signal integrated circuits. Following contents are covered in the dissertation:;(1) A general stimulus error identification and removal (SEIR) algorithm that can test high-resolution ADCs using two low-linearity signals with a constant offset in between; (2) A center-symmetric interleaving (CSI) strategy for generating test signals to be used with the SEIR algorithm; (3) An architecture-based test algorithm for high-performance pipelined or cyclic ADCs using a single nonlinear stimulus; (4) Using Kalman Filter to improve the efficiency of ADC testing; and (5) A testing algorithm for high-speed high-resolution DACs using low-resolution ADCs with dithering.