Meaning
Statistical analysis evaluates the hierarchical sources of variability across different production runs to ensure process consistency. Utilizing anova nested variance allows manufacturing engineers to separate the variation caused by component batch changes from the variation introduced by individual assembly stations. This technique restricts its domain to nested designs and cannot be applied to crossed experimental structures where every level of one factor occurs with every level of another.
Hierarchical Allocation
In a nested model, each sub-factor exists exclusively within a specific higher-level category. The nested structure might place individual printed circuit board panels within specific raw material lots, allowing a precise isolation of wafer-level and panel-level effects. By partitioning the total sum of squares into nested components, anova nested variance identifies which level of the manufacturing hierarchy dominates the inconsistency.
This quantification guides equipment tuning by highlighting whether the primary source of error lies within the component placement machine or the chemical bath.
Variance Apportionment
Calculations yield distinct variance components that sum to the total observed variation in the product. These values show the percentage of instability attributable to each hierarchical stage. This representation is critical when determining the source of thermal drift in RF modules across multiple shifts.
Process Control
Stability thresholds are derived from the calculation to prevent unnecessary assembly line stops. High nested variance at the wafer level indicates that adjusting the pick-and-place machine would be ineffective. Corrective actions are directed toward the supplier material when the wafer level shows the highest drift.
