Capability measurement analysis for destructive testing by nested design and improve process using taguchi DOE

Abstract

Liquid particle counting (LPC) is used widely in hard disk drive industrial manufacturing. The capability
measurement analysis or GR&R is used to evaluate capability of LPC test method. Because the LPC test
method is the destructive testing, so the nested design is used in GR&R study. The results of GR&R study
show that the GR&R is 81.35% and unacceptable because it exceeded 30% according to the AIAG measurement
system analysis criteria. The result of experiment showed that this measurement system must be improved.
First, the cause and effect diagram is used to identify the sources of measurement variation. There are 4 main
sources of variation; method, environment, part and the different of operator. Finally two main sources of
variation (operator and method) are improved by Taguchi DOE. In Taguchi DOE the control factor are the
curling time, the position of sample, and the noise factor which is the different of operator. The optimal
conditions from Taguchi DOE is curling time be equal to 4 hr and the position of sample is bottom. Furthermore
the noise factor is not effective in the measurement process.