Using JMP® to Automate Capacity Model Input Derivation in the Semiconductor Manufacturing Industry
Aug 12, 2015 1:00 PM
Jason Kintz, Senior Industrial Engineer, ON Semiconductor
A capacity model is an invaluable tool for driving manufacturing decisions, and its accuracy is directly affected by its inputs. As an industrial engineer at ON Semiconductor, I am responsible for the derivation and accuracy of more than 8,000 tool- and recipe-level model inputs. I searched for software to automate this process, and found JMP to be the clear standout among the various packages available. The approachable scripting language, intuitive user interface and extensive online resources allowed me, a first-time JMP user, to bring my data automation ideas to fruition in a short period of time. I created several JSL scripts to save our company time and money. A simple five-step approach was used: Connect to an Oracle database, extract and sort the data required for calculations, perform calculations, screen the data for outliers, and generate summary tables and reports. The scripts generate recipe-level process time, setup time and a measure of how well the operators loaded the tools. Additional benefits include a tool matching report, an outlier summary and a from-to delay matrix we use in our dispatching logic to minimize recipe changeover time. ON Semiconductor has been using the scripts at its Oregon site for more than a year, and this process was selected as a Global Industrial Engineering initiative for 2015. The company plans to roll out the scripts to other manufacturing sites in Idaho, Belgium, Czech Republic, Japan and Malaysia.