FDC Lite is a lite version of standard FDC System. Use Web form user interface and user do not need to build MVA(Mutiple Variable Analysis) model in advance. Perform temporal Data monitoring and abnormality-detection directly. Alarm is sent once there is an abnormality, and actions (such as: make the tool out of service or notify the engineers…, etc) are taken to avoid continuing product-abnormality.
|System Difference||FDC Standard||FDC Lite|
|Features||Mutiple detect module, compare the eigenvalues of each variable comprehensively and provide Global Health Index(GHI).||FDC cover rate can reach to 100% and fully monitor temporal data.|
|Applicable Process Features||Apply in complicated process equipments(e.g. the number of recipe, step and sensor is big and will effect each other, and doesn't use mean as the index of indicator). Applicable in large-volume, small-variety mass production.||Apply in simple process equipments(e.g. the number of recipe, step and sensor is small and always convert data into indicator by mean). Applicable in small-volume, large-variety production.|
|Application Process||IC FEOL, LCD Production Process||Packaging, MOCVD, LED BEOL, Netcom Industry|
|Model Base||By Recipe(/By Chamber)||By Tool / By Recipe / By Chamber|
|FDC Model Difference||Convert mutiple variable into GHI Model ang monitor, also monitoring single variable||Monitoring single variable|
|Mutiple Variable Analysis(MVA)||Top 10 abnormal variable analysis & sort(ISI Analysis), Notification of single variable abnormal||Notification of single variable abnormal|
|Online Learning||Tuning(auto learning) mutiple variable model characteristic online||Tuning single variable spec online|
Features & Functions
System will alarm when tool abnormal and inform engineer take action to avoid continuing production of defective products.
Establish Monitor Rule: Perform sensor spec setting(By Manual, Sigma, Percentage) and data quality filter.
Auto rule setting and monitoring by "General Rule" once there is new production be detected.
Real time monitoring tool sensor by using Temporal, Mean, Max, Min, Range, Std, Slope, Counter, Accumulation alogrithm..., etc.
Report: Classify data by time period, tool type, recipe and provide OOS/OOC statistic chart(by piece, daily, weekly, monthly).
Specific tool/recipe "Alarm Rate" setting. When alarm rate exceed the standard, tool/recipe will be marked in red for convenient management.
Provide OOS sensor list of each piece/run. One click to check temporal data of front 3 pieces, helping engineer find root cause of defect qulckly.