Commonality Analysis refers to a set of techniques used to identify systematic causes of yield loss. Because of the complexity and number of process steps involved in semiconductor manufacturing, pinpointing the cause of a yield issue can be a challenging task. yieldWerx Enterprise enables product engineers to systematically evaluate common variables within manufacturing processes to efficiently identify the source of yield issues.
Product engineers verify chip design robustness and document the variance of important electrical measurements for a new IC or process through device characterization. Characterization analysis screens for problems to ensure adequate yield for target specifications. Primarily performed on “split lots,” device characterization results are summarized in a characterization report that includes Cpk (process capability index), yield forecast data, and correlation analysis.
Gauge R&R studies are critical to identifying unknown issues in measurement quality, and ANOVA is the preferred method for analyzing Gauge R& R results. ANOVA calculations are both highly detailed and time-consuming, creating a need for tools to automate the complex calculation process. yieldWerx Enterprise provides ANOVA tools for stakeholders concerned with measurement quality throughout the enterprise, including product, test and manufacturing engineers.
One of the most challenging aspects of yield management is ensuring that data is consistent and accurate. Sifting through large volumes of data is a time-consuming task for engineers. yieldWerx Enterprise’s Smart Logic data cleansing engine addresses data quality issues by mapping and correcting critical fields that are missing or inaccurate as data is being loaded into the warehouse.
Because product costs are added with every step in the semiconductor supply chain, eliminating low-yielding material early in the manufacturing process optimizes overall costs. yieldWerx Enterprise can correlate Final Test yields to Wafer Probe yields to eliminate bad material at probe, leading to package, assembly and test cost savings.
As standard wafer sizes have grown, the variation in electrical performance across a wafer has increased. Averaging the electrical results from a typical five-site Process Control Monitor (PCM) has become inaccurate. To improve the accuracy of this analysis, yieldWerx Enterprise allows you to break a wafer into “zones” and use the PCM data from the nearest zone for correlation analysis and more.