Early Detection with OT-Detect
In a global operation environment, subcontractors or remote facilities are accountable for the operation’s performance and health. Optimal+'s Early Detection capability, now with Optimal+-Detect, allows Fabless organizations to track performance metrics more effectively and to ensure maximum values for yield, quality, utilization and throughput are achieved. Optimal+-Detect, an innovative enabling technology, automatically tracks ALL products for ANY changes in baseline production. This means that any extreme change in your products' manufacturing, test or assembly processes will be tracked, captured and assessed.
The Early Detection module consists of a number of components and capabilities:
- Optimal+-Detect develops a baseline of products' yields, bins, soft-bins, failing parameters and more. Hundreds of baselines can be developed for dozens, hundreds, or thousands of products, each with all of that done on the fly. Once triggered, it provides a step-by-step root-cause analysis.
- OT-Proxy is deployed on testers and collects data in real time
- Once testing of a wafer or lot is complete, an Optimal+DF datalog is sent securely and automatically to the OT-DB, typically located at the Fabless HQ
- Data is cleansed and augmented automatically to ensure rich and consistent data from all suppliers
- On arrival (usually within minutes of the end of test), data is immediately available for viewing and analysis in OT-Portal
- With OT-Rules, users can define Statistical Process Control (SPC) rules which execute automatically when data arrives and send email alerts if issues are identified. Rules can use dynamically calculated or pre-specified baselines and cross-fleet performance comparisons to catch problems early
- Alerts triggered by OT-Rules include an attachment with detailed analysis data that can be viewed and further analyzed in OT-Portal.
Benefits [read more]
- Track, capture and assess any extreme change in your products' manufacturing, test or assembly
- Immediate "plug and play" usage with structured methodology provided with Optimal+-Detect
- Catch issues at any test house in near real-time so they can be resolved quickly and with minimal impact
- Find root causes of issues with the help of comprehensive and powerful analysis tools linked directly to the data
- Save time by defining rules that pinpoint problems for you
All of these benefits will increase yield, throughput and ATE utilization and improve quality
Additional benefits for Fabless organizations with engineering test floors, NPI labs or consigned testers, are described under the Test Floor Operations solution section.
OT-Rules Early Detection Rules Categories [read more]
The following rules are available in OT-Rules to process data as it arrives shortly after testing ends:
- Yield Monitor – identifies lots with abnormal yield
- Statistical Bin Limit (SBL) – looks for abnormal levels of specific hard/soft bins
- Site-to-Site Bin Deviation – used in parallel testing to identify a test-site with a significantly higher specific hard/soft bin signature than other test-sites
- Site-to-Site YieldDeviation – used in parallel testing to identify a test-site with significantly lower yield than other test-sites
- Site-to-Site Fail Test Deviation – used in parallel testing to identify a test-site with parametric tests showing statistically different behavior than other test-sites
- Fail Test Limit – identifies wafers or lots with abnormal failures of specific parametric or functional tests
- Parametric Test Trend – captures a trend in the combined data from one or more parametric test and alerts the user before the results move out of limits causing parts to fail
- Parametric Test Freeze – detects situations where due to a software or hardware failure some of the tests measurements are discontinued and the data log file has several consecutive die with the same or similar test values on a specific parametric test
- Parametric Process Capability – identifies a parametric test that exhibits a wide spread or a degradation in process capability on a specific wafer or lot
- Outlier Detection – a comprehensive set of rules to identify and re-bin marginal parts
- TTR Monitor – part of the TTR Solution which monitors lots on which TTR was performed to detect issues with the TTR process such as those caused by unexpected test program changes
- New in 5.0: Generic Data Rule - monitors any measure available in OT-Portal and reports when a specific lot has a result which is significantly different from a dynamically calculated baseline. This rule can be used to track a wide variety of parameters such as test time, first-pass yield, retest rates, etc.
- New in 5.0: Extended Site-to-Site - signals intermittent hardware failures on a test site level in the case of high-parallelism testing. These failures can be found by comparing the data bin failure rates at each test site and applying a statistical methodology to determine if some sites are failing significantly more often than others. A novel feature of the algorithm is that it automatically performs the assessment of a potential yield loss by all malfunctioning test sites, and the alarms are triggered only when the cumulative yield loss exceeds the target.
Features [read more]
- Actionable data available almost immediately (in a matter of minutes)
- Perform baselines by automatically scanning for a specific parameter to track and assess. For example, a soft-bin behavior across the last 20 - 40 lots, and check if the lot's current soft-bin signature statistically exceeds that baseline value by X sigma or percent, etc.
- Perform many hundreds, and sometimes thousands, of baselines on the fly for each incoming wafer, run or lot, using optimized algorithms for super-fast computations.
- Near-time and offline predefined rules for yield loss detection, quality issues and throughput impact
- Email notifications to a distribution list both "in" and "out" of your organization, including supply chain operation engineers and managers
- Offline simulation of new rules and evaluation of execution results in a controlled environment, prior to activation in production
- Data analysis with drill-down capabilities to identify potential operational issues and/or potential improvements using root cause analysis methods
Examples [read more]
Click on the examples below to see how Optimal+ tools enabled engineers and managers to find and fix issues easily and quickly.
- Maximizing Yield through Probecard Performance Analysis
- Using Rules to Catch ATE Issues
- Uncovering Geographic Issues at Final Test (Data Feed Backward)
- Analyzing Test Program Performance
- Preventing Test Time Degradation
- Comparing Test Time Across a Fleet of Testers
- Managing Online Retest Settings
- Reducing Wasteful Retests