Test Time Reduction and Quality Augmentation
The Test Time Reduction (TTR) and Quality Augmentation Solution consists of a powerful combination of offline simulators and analysis tools, coupled with real-time tester components that interact with the test program to monitor test results and dynamically switch tests on and off.
It may seem strange to you that we have combined TTR and Quality Augmentation into a single solution. The reasoning behind this is that they can both be achieved using an identical process. Here's how.
In adaptive TTR, existing tests are "switched off" dynamically by OT-Proxy when a TTR "rule" deems that it is safe to do so. In order to maintain the quality of the TTR process these tests are "switched on" again every few touchdowns and the result is sampled to check that the test is still within control limits.
For non-TTR-related quality, it is possible to use the same mechanism to "augment" test data by switching expensive tests on for a small number of sampled tests, enabling engineers to collect more detailed test data with less test time impact than previously possible.
- Tests which fail after a bin was assigned in continue-on-fail scenarios
- Tests which also fail in good parts
- Tests which fail but belong to a test suite which eventually passes
Test Time Reduction (TTR) is a process in which test time is saved by disabling tests which are typically causing little or no fallout, thus saving test time. There are two flavors of TTR supported by the OptimalTest suite of tools:
- Classical TTR – where simulation tools enable users to determine which tests can be deleted from a test program by analyzing historical data to find tests that rarely or never fail.
- Advanced Adaptive TTR – where the results of the simulation process are used to generate a TTR “rule” which controls the test program and enables or disables tests based on sampled test results. Since the "removed" tests are still sampled, higher quality is achieved and it is possible to stop performing TTR when a maverick lot is detected.
Quality Augmentation uses the same online capabilities provided by the Advanced Adaptive TTR solution to enable long-running tests to execute only on a small sample of parts. This is a great way to run tests to capture additional test data which would be too costly if executed on every part.
OptimalTest's unique TTR and Quality Augmentation solutions provide a complete end-to-end solution for TTR that can work across a distributed supply chain. A typical TTR process consists of the following steps, all of which are supported and automated by OptimalTest tools:
- Running an analysis report to identify potential candidates for TTR
- Simulating a TTR process based on the potential candidates on historical data in order to estimate the amount of time saved and potential impact on Defects per Million (DPM)
- (Optionally) performing classical TTR by deleting tests that are considered redundant
- Publishing an Advanced Adaptive TTR rule to the test location (which could be another company in the supply chain for fabless companies or a different location in an IDM) through the automatic rule publication and monitoring system in OT-Rules
- For wafer sort, optionally changing the probing sequence to probe strategically located “reference dice” with all tests enabled before probing the rest of the wafer
- Running Advanced Adaptive TTR in production
- Collecting detailed TTR feedback from the production processes for analysis in OT-Portal
- Monitoring the entire process for issues such as test program changes that might impact TTR
OptimalTest TTR offers a selection of three algorithms:
- Sampling TTR - the rule performs straightforward sampling of tests based on a constant sampling rate (for example, performing TTR on 9 out of 10 touchdowns, and full test flow on every 10th touchdown). This is also the methodology used for Quality Augmentation.
- Pass/Fail-based TTR - the rule performs sampling of each candidate test and based on the number of fails, makes a real-time decision on execution of TTR on the rest of the parts.
- Adaptive Parametric-based TTR - an advanced algorithm with the prediction of the parametric test range based on sampling test results. If the predicted test limit is approaching (but not exceeding) critical values close to test spec limits, TTR can be automatically discontinued. Additionally, Adaptive TTR has a mechanism of online re-evaluation of the test range so if after several more touchdowns the predicted test range returns to a safe area, TTR can be restarted.
OptimalTest TTR is geography aware and can apply varying levels of TTR to different areas of a wafer. For example, the edge of a wafer may require more intensive testing than the center of the wafer.
Additionally, OptimalTest is currently developing additional TTR algorithms which will leverage the data in OT-DB to provide even more control over the TTR process:
- TTR based on ULT (Unit Level Traceability) where data from previous operations is used to determine how much testing is required in the current testing
- TTR based on E-Test data where E-Test data is used to control the amount of testing the part requires
- Save test time without compromising quality by using adaptive test time reduction
- Augment your test process with tests usually considered too expensive by using the sampling capabilities of the OT-Proxy TTR engine
- Ensure the validity of your TTR recipes by analyzing and simulating TTR rules using historical test data
- Avoid modification to test programs by leveraging the OptimalTest template-based rule mechanism to define TTR rules
- Keep control of the process by publishing rules across your supply chain automatically and tracking the state of your TTR activities from one central location in OT-Rules
- Catch TTR-related test program issues early by implementing TTR monitoring rules which alert when too much or too little TTR was executed for a lot
- Tune your TTR rules to perfection by reviewing the TTR feedback data collected automatically into the OT-DB
- Debug TTR rules on the tester to automatically and exhaustively check that combinations of disabled tests don't break the test program.
- Powerful wafer sort and final test analysis and simulation tools for TTR
- Rich set of TTR algorithms
- Automatic TTR rule publication
- Support for geographic TTR in wafer sort
- Template-based definition of TTR rules
- Monitoring rules to track the quality of the TTR process
- Detailed TTR feedback down to the individual part and test level (which tests did or didn’t execute and why)
- Fully integrated into the OptimalTest suite of tools and applications
- NEW with 5.0: Flexible sampling rates and numbers of validation units for different test suites
- NEW with 5.0: Increased TTR opportunities achieved by automatically compensating for non-critical tests which fail when Continue-on-Fail (CoF) is enabled
Additional Test Time Reduction and Quality Augmentation Modules
The following modules also support the TTR solution:
More information about using Adaptive Test Time Reduction in production can be found in joint presentations made by Qualcomm and OptimalTest at the Teradyne Users Conference (TUG 2011) and the Verigy Users Conference (VOICE 2011).