To start off the day, I was instructed to complete a partial report containing just what I had learned during the first day. Some of the details that I had written are Assembly Plant Malaysia (APM) processes, introduction to semiconductor and organizational chart.
Upon completion of the first task, my supervisor, Mr. Melvin presented me with a mathematical problem involving wire bonding process. I was asked to find the percentage increase in gold wire cost if each individual die size is reduced by 4% from the two edges. Note that the die height is still the same and relevant parameter values were given according to their engineering specifications. So, by using the famous Pythagoras theorem, I was able to calculate the percentage increase in cost. Then, I was also required to devise a general equation for evaluating the percentage cost increase which consists of four variables, namely wire bonding horizontal length, percentage increase wire bonding horizontal length, die height and line height.
After the brief lunch break, I was introduced to DOE which stands for Design of Experiment as well as SPC (Statistical Process Control). In DOE, I had gained some basic knowledge on how to implement 2k full factorial design. It is unequivocal that for each process, there may be several input variables. Through DOE, and via Minitab software, I was able to produce information regarding the set of input variables as well as the output variables. There are four models of interest. Full model, final reduced model, graphical analysis and mathematical model. I was able to determine which input variables are statistically significant or otherwise. Those insignificant factors will be neglected for further analysis of the operation. At last, one would be able to determine the equation that governs the relationship between the sole output and a set of input variables. Besides that, the determination of process window, which can be esteemed as the range of allowable process variables, is of utmost importance. Process operation should be based on the values that are at the center of this particular process window.
In SPC, I was exposed to the use of process capability index, or capability index. This index relates the voice of the customer (specification limits) to the voice of the process. In other words, it is the ratio of the distance from the process center to the nearest specification limit divided by a measure of the process variability. This parameter is useful in determining whether the process if stable or otherwise as it inadvertently reduces the complex information about the process to a single number. If the result were to be greater than 1.67, then the process is reckoned to be stable. Customarily, we deal with two limits, one being the upper sided limit, while the other one, naturally, being the lower sided limit. So, this inextricably means that there will be two calculations of the capability index. The one which yields the lowest value is the one we are going to compare it with the standard value of 1.67. As an extra information, process capability index is influenced by the way the process data are collected, that is, the process view. Process view can be defined in terms of the time frame, sampling method (sampling frequency, size...) used to obtain process data.
No comments:
Post a Comment