Die-level adaptive test: Real-time test reordering and elimination

Test data analytics — Exploring spatial and test-item correlations in production test data

“…The values of the corresponding candidate test items can be predicted by a combination of other test items. Since we use a penalty parameter, λ, in the minimization problem (17) to control the sparse level of α, each u of the solution found may not be minimal. That is, there is no guarantee on the prediction accuracy for the candidate test items derived from the correlation model.…”

“…An adaptive test flow was proposed in [16] that adjusts the test process to devote more test resources to marginal devices and less resources to passing devices with large margins. In [17], an adaptive test scheme was proposed to dynamically control the test flows and test contents on-tester in realtime through continuous per-die updates of test fail rate.…”

“…Known applications in production testing that utilize the intertest-item correlations include: an early work that used a Monte Carlo based approach to analyze the joint probability distributions of test items for constructing a regression model of the untested performances [23], reordering the test items to screen out faulty dies earlier [13], and predicting passed test items to optimize the test set accordingly [16], [17], [24].…”

A test-application-count based learning technique for test time reduction

“…In [3], the authors proposed a die-level test reordering and elimination technique for TTR (test time reduction). The test reordering technique keeps track of, for each pattern, the number of times it is applied and the number of faulty dies it detects, denoted by β and α, respectively.…”

“…In [3], it is observed that test pattern reordering mainly occurs in the front of the test list, i.e., the high detection ratio test patterns, and the patterns at the end rarely change their positions or remain stable. (This is also reported in [2].)…”