| This thesis focuses on a scan-based delay testing technique that was recently developed at ETS. This new approach, called Captureless Delay Testing (CDT), has been proposed as a technique that complements traditional methods of test, ensuring the integrated circuits will function at their proposed clock speed, further improving the test coverage of the particular type of test. Furthermore, CDT incorporates the use of sensors enabling the detection of the presence of transitions at strategic locations.;Secondly, we propose a fully automated algorithm that enables, at the earliest stages of the test vectors generation process: 1) the identification of the non-covered nodes, 2) the identification of the placements of the CDT sensors at the inputs of the flip-flops for further improvement of the test coverage, and 3) the minimization of the number of sensors with regards to requirements. Our results indicate that when we apply CDT on top of transition-based fault model we can improve the test coverage by 5%. Moreover, the algorithm of CDT sensors minimization allows a reduction of more than 85% the number of those sensors with a minimal test coverage loss, on average of 1.6%.;Keywords: analogue circuits, automatic test pattern generation, captureless delay testing, integrated circuit testing, low cost testing, scan-based test technique;The purpose of this project is to improve on certain aspects of this novel technique. At first, we analyze the delay distribution of the non-covered nodes by traditional methods of test, in order to develop the best way possible of placement of the CDT sensors. We present, using Perl Language, the. ensemble of tools developed for this purpose. The end results obtained confirm that the paths that pass through the non-covered nodes are longer than those that traverse the covered ones. The difference between the two types of paths exceeds 20% of the clock period when considering the shorter path delay values. |
