Page 88 - Six Sigma for electronics design and manufacturing
P. 88

item, the specification limits are 10.00	 ± 0.04	. The process is charac-
                        terized by   = 10.00	 and   = 0.015.
                          Cpk = Cp = ± SL/± 3   = 0.04/3 · 0.015 = 0.89
                          z = 3 · Cp = 2.67; f(–z) = .0038; RR = 0.38%; for each side above the
                            USL and below the LSL. Percent OK = 1 – total RR = 99.24%.
                         Note that in this case, the Cp was less than 1, therefore it is expect-
                        ed that the reject rate would be higher than 3   design defect rates of
                        2700 PPM or 0.27%.   The Elements of Six Sigma and Their Determination  57
                         For  the  second  item,  the  specification  limits  remain  the  same  at
                        10.00	 ± 0.04	. The process is shifted from the first item by 0.01	 and
                        characterized by   = 9.99	 and   = 0.015.
                          Cp = 0.89 (remains the same from the first item)
                          Cpk = min(USL – average/3 ) or (average – LSL/3 )
                          Cpk = 0.67 = minimum of (10.04 – 9.99/3 · 0.015) = 1.11
                          or (9.99 – 9.96/3 · 0.015) = 0.67
                          z 1 = 3 · Cpk (low) = 3 · 0.67 = 2.00; f(–2) = 0.0228 or 2.28%
                          z 2 = 3 · Cpk (high) 3 · 1.11 = 3.33; f(–3.33) = 0.00043 or 0.43%
                          Total RR = f(–z 1 ) + f(–z 2 ) =  0.00228 + 0.00043 = 0.02323
                          Percent OK = 1 – total RR = 1 – 0.2323 = 0.97677 or 97.68%.
                         It  is  apparent  that  if  the  manufacturing  process  is  not  centered
                        with the specification nominal (second case in the table), the total de-
                        fect rate increases, even if the manufacturing process standard devia-
                        tion remains the same. Similar increases in the defect rate occur if the
                        manufacturing process standard deviation increases or there is a com-
                        parable decrease in the tolerance limits of the design. The table also
                        illustrates the use of Cp or Cpk as indicators of quality, depending on
                        whether  the  manufacturing  process  average  is  equal  to  the  design
                        specification nominal.

                        2.3.3  Attribute processes and reject analysis
                        for six sigma
                        For attribute processes (those with quality measured in terms of de-
                        fects in a sample or number defective), an implied Cpk will have to be
                        calculated in the quality assessment of design and manufacturing. It
                        is assumed that defects are occurring because of violation of a particu-
                        lar or a composite specification(s). The composite specification can be
                        one-sided or two-sided, depending on the interpretation of the defects.
                        For example, a wire bond defect could be the result of one-sided speci-
   83   84   85   86   87   88   89   90   91   92   93