Background of the Project
Pokamisu occurences in Pre-operation calibrations of Test Machines have become a common concern accross several semiconductor products, for operators who had grew confident in performing this repeatitive task. It is human tendency that overtime, these repeatitive tasks become nothing more than a hand and pen movement activity. Operators become incognizant of the purpose of the activity, validity of data and accuracy of the calculations. Pre-operation calibration is a necessity, thus it is not an option to remove it from the process sequence or reduce the already established frequency as any changes in such factors will put the reliability of the test process in crucial, unforseen deterioration then failure.
Common pokamisu ranges from using the wrong calibration samples, errorneous transfer of or reference to standard values, wrong calculations to wrong conclusions whether the test machine had passed or not the calibration. A simple cause and effect analysis on these failure modes, puts operators' confidence ( actually boredom, lack of interest ) and frequent manual data transfers/calculations as primary causes.
Having identified the primary causes of pokamisu, the best corrective action then for this problem is a pre-operation calibration system that is independent of operators' level of interest in the activity and reduced if not elliminated manual data transfers or referencing. A software that is, that will do all the calculations, data readings and data transfers and finally dictate, and not suggest, the results of the calibration.
Common pokamisu ranges from using the wrong calibration samples, errorneous transfer of or reference to standard values, wrong calculations to wrong conclusions whether the test machine had passed or not the calibration. A simple cause and effect analysis on these failure modes, puts operators' confidence ( actually boredom, lack of interest ) and frequent manual data transfers/calculations as primary causes.
Having identified the primary causes of pokamisu, the best corrective action then for this problem is a pre-operation calibration system that is independent of operators' level of interest in the activity and reduced if not elliminated manual data transfers or referencing. A software that is, that will do all the calculations, data readings and data transfers and finally dictate, and not suggest, the results of the calibration.
Objective of the Project
To develope a test machine calibration software.
Through this software, reliability of the pre-operation calibration will become independent of the operators' level of interest on the task and will remove majority of manual data transfer/referencing, as well as elliminate hand calculations. Decision making is likewise removed from the operator as this software will provide calibration results as passed or failure.
Through this software, reliability of the pre-operation calibration will become independent of the operators' level of interest on the task and will remove majority of manual data transfer/referencing, as well as elliminate hand calculations. Decision making is likewise removed from the operator as this software will provide calibration results as passed or failure.
Methodology
1.Language - VBA for Excel.
All calibration methods require standard values to serve as baselines and since test machines can be utilized to test several models or type series of a semiconductor product, this would also mean several, different sets of standard values. With VBA for Excel, standard values can be easily added as worksheets and protected. Also we take advantage of the readily available graphical decisions tools ( graphs ), where its usefulness will become apparent in the final output of the software.
2. Data and Results
All calibration methods require standard values to serve as baselines and since test machines can be utilized to test several models or type series of a semiconductor product, this would also mean several, different sets of standard values. With VBA for Excel, standard values can be easily added as worksheets and protected. Also we take advantage of the readily available graphical decisions tools ( graphs ), where its usefulness will become apparent in the final output of the software.
2. Data and Results
Raw data (the .dat file) will be accessed directly from the hard drive of the test machine but will not be modified and no duplicate copies will be made, unless calibration will be run in another computer. Standard values as said, will be accessed but will be password protected. Summary of results will be contained in a dialog box as final output of the calibration. Calculated values together with the date and time will be compiled in a single file and will be saved in a secondary hard drive ( D: \ ) of .csv format. Compiling the results in a single file is important as this will be useful in performance analyses of the test machine over a period of time.
3. Implementation
MS Excel will be installed in the tester ( the desktop computer attached to the handler) and the calibration software will be run thereon. Both the Software calibration and Manual Calibration will have to be run in parallel during the first few weeks of the implementation to ensure correct results for the software.
MS Excel will be installed in the tester ( the desktop computer attached to the handler) and the calibration software will be run thereon. Both the Software calibration and Manual Calibration will have to be run in parallel during the first few weeks of the implementation to ensure correct results for the software.
Output and Code
Output dialog box presents an effortless, thoughtless way of knowing whether the machine passed or failed the calibration. Even without reading, a glance is enough to know the calibration status. Seeing a shade of red in the Calibration Status group is all that is necessary to confirm calibration failure, otherwise no further thought is needed, production can proceed normally.
In the case of failure, one can easily trace the cause of failure through inspection of the failed calibration component - the red ones. Further, identifying which of the calibration units failed is easy through the graph displayed where the standard values on these graphs are fitted with maximum and minimum error bars for easy visual identification of units failed. Switch graphs to display by selection on the radio buttons.
