TABLE OF CONTENTS Experiments allow investigators to determine the effects of several variables on a given process or product. A designed experiment is a test, or series of tests, which involve purposeful changes to these variables in order to observe the effect of the changes on that process or product. Statistical experimental design is an efficient approach for systematically varying these controllable process variables, and ultimately determining their impact on process/product quality. This approach is useful for comparing methods, deducing dependencies, and creating models to predict effects.
Statistical process control (SPC) and experimental design are closely interrelated. Both techniques can be used to reduce variability. However, SPC is a passive approach in which a process is monitored and data is collected, whereas experimental design requires active intervention in performing tests on the process under different conditions. Experimental design can also be beneficial in implementing SPC, since designed experiments may help to identify the most influential process variables, as well as their optimum settings.
Overall, experimental design is a powerful engineering tool for improving a manufacturing process. Application of experimental design techniques can lead to:
Improved yield
Reduced variability
Reduced development time
Reduced cost
Ultimately, the result is enhanced manufacturability, performance, and product reliability. The following sections illustrate the use of experimental design methods in electronic packaging manufacturing.
Consider the yield data in Table 20.1 obtained from a PWB manufacturing process in which two batches of ten printed wiring boards were fabricated,...
