Deterministic dynamic element matching: an enabling technology for soc built-in-self-test

Hanjun Jiang,Degang Chen
DOI: https://doi.org/10.31274/rtd-180813-15384
2005-01-01
Abstract:The analog-to-digital converter (ADC) is a key building block of today's high-volume systems-on-a-chip (SoCs). Built-in-self-test (BIST) is the most promising solution to testing deeply-embedded ADCs. Cost-effective stimulus source with on-chip integrability has been viewed as the bottleneck of ADC BIST, and consequentially the bottleneck of SoC BIST and BIST-based self-calibration. The deterministic dynamic element matching (DDEM) technique has been proposed as a solution to this problem. In this work, rigorous theoretical analysis is presented to show the performance of a DDEM digital-to-analog converter (DAC) as an ADC linearity test stimulus source. Guided by the insight obtained this analysis, a systematic approach for cost-effective DDEM DAC design is proposed. Two generations of DDEM DACs have been designed, fabricated, and measured. 12-bit equivalent linearity was achieved from the first DDEM DAC with 8-bit apparent resolution and less than 5-bit raw linearity after systematic error compensation. The achieved 12-bit linearity outperforms any on-chip stimulus source in literature. Based on the first design, a new DDEM DAC with 12-bit apparent resolution, 10-bit raw linearity, and 9-bit DDEM switching was designed with improved design technique. This DAC was fabricated in standard 0.5-pm CMOS technology with a core die area of 2 mm2. Clear ramp signals could be observed on an oscilloscope when the DDEM DAC was clocked at 100 MHz. Laboratory testing results confirmed that the new DDEM DAC achieved at least a 16-bit equivalent linearity; this was limited by the available instrumentation, which has 18-bit linearity. It outperforms any previously reported on-chip stimulus source in terms of ADC BIST performance by 5 bits. The robust performance, low cost, and short design cycle for on-chip implementation make DDEM an enabling technology for SoC BIST and self-calibration. Two new approaches based on DDEM are developed to further boost the die area efficiency, improving the basic DDEM approach. The first is termed segmented DDEM, and the second is dither-incorporated DDEM (DiDDEM). It has been shown through mathematical analysis and simulation that these can maintain the performance of the basic DDEM approach while greatly reducing the implementation cost.
What problem does this paper attempt to address?