ADC-Based Power Droop Testing for Pattern-Level Fault Isolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing circuit testing methods face challenges in identifying specific test stimulus patterns that cause power anomalies, such as power droop or spike, which can disrupt circuit operation and require costly redesigns or resets, due to the difficulty in determining which portion of a test pattern results in the largest voltage deviation.
Innovation Solution
A test circuit system that identifies a particular test stimulus pattern causing power anomalies by sampling voltage levels during test applications, re-applies the pattern with varying start times to isolate the portion responsible for the anomaly, using an ADC circuit to detect minimum voltage levels and store samples for analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional testing methods are used to identify power anomalies, then testing can be performed, but the specific portion of the test pattern causing the anomaly cannot be precisely identified
Solution Approach 1:
The test pattern is divided into multiple segments by applying it repeatedly with different start times. Each application focuses on a different portion of the pattern, allowing systematic identification of the specific segment causing the power anomaly. This segmentation transforms an unidentifiable whole-pattern problem into discrete, analyzable portions.
Solution Approach 2:
The voltage sampling is performed at predetermined times relative to the test pattern start time. By establishing fixed sampling points before knowing the exact anomaly location, the system prepares measurement data that will reveal which pattern portion causes the anomaly. This preliminary positioning of measurements enables subsequent precise identification.
2Measurement precision
If voltage sampling is performed at fixed intervals, then measurement is simple, but the sampling may not capture the exact moment of power anomaly
Solution Approach 1:
The test pattern is applied periodically multiple times with different start times. Each periodic application shifts the sampling window to capture different portions of the pattern. This periodic variation in start time ensures that the power anomaly moment is captured in at least one of the repetitions, improving detection accuracy without requiring continuous monitoring.
3Measurement precision
If the entire test pattern is re-applied multiple times, then the anomaly source can be isolated, but test time increases
Solution Approach 1:
Instead of analyzing the entire test pattern at once, the method applies the pattern multiple times with varying start times, effectively examining only the relevant portions repeatedly. This partial action approach focuses test resources on identifying the specific anomaly-causing segment rather than uniformly testing all pattern portions, reducing overall test time while maintaining precision.
Data Source
AI summary
An apparatus includes a functional circuit, including a power supply node, and a test circuit. The functional circuit is configured to operate in a test mode that includes generating respective test output patterns in response to application of a plurality of test stimulus patterns. The test circuit is configured to identify a particular test stimulus pattern of the plurality of test stimulus patterns, and to reapply the particular test stimulus pattern to the functional circuit multiple times. The test circuit is further configured to vary, for each reapplication, a start time of the particular test stimulus pattern in relation to when a voltage level of the power supply node is sampled for that reapplication.