Product Manufacturing CHEN 4470 – Process Design Practice Dr. Mario Richard Eden Department of Chemical Engineering Auburn University Lecture No. 11 – Introduction to Six Sigma in Product Manufacturing March 6, 2007 Contains Material Developed by Dr. Daniel R. Lewin, Technion, Israel
Be able to define the Sigma Level of a manufacturing process
Know the steps followed in product design and manufacture (DMAIC)
Be able to qualitatively analyze a process for the manufacture of a product and know how to identify the CTQ step using DMAIC
Example: The Electronics Food Chain
Product Development Source: Dataquest 1999 data Electronic Equipment and Systems $988 B Semiconductors $170 B Materials and $33 B Equipment Semiconductor
IC Production Capability
Device Complexity Trends
Product Development Chip Area Device Year Transistors per Chip (cm 2 ) 8086 1978 30K 0.34 80286 1981 120K 0.77 80386 1985 400K 1.0 486 1990 2M 1.8 Pentium 1993 3.5M 2.9 Pentium Pro 1995 5.5M 2.9
Technology vs. Economics
Product Development Physical Limit Economic Limit The budget always runs out before the physical limits are reached. Capability Cost
Technology vs. Economics (Continued)
Product Development Physical Limit Economic Limit Capability Cost New Physical Limit Innovation!!
Implications of Blind Faith in Moore’s Law
Fear is that exponential growth is only the first half of an “S” shaped curve
Product Development Revenue Time
Industry Drivers (Push vs. Pull)
Market requires (push):
Smaller feature sizes desired
Larger chip area desired
Improved IC designs lead to innovations
IC industry delivers (pull):
Lower cost per function (higher performance per cost)
New applications are enabled to use chips with new capabilities
Higher volumes produced
6 = “Six Sigma”
SSL: Chapter 19
Structured methodology for eliminating defects, and hence, improving product quality in manufacturing and services.
Aims at identifying and reducing the variance in product quality, and involves a combination of statistical quality control, data analysis methods, and the training of personnel.
Six Sigma 1:15
is the standard deviation (SD) of the value of a quality variable, x , a measure of its variance, assumed to be normally distributed:
Assume L ower C ontrol L imit LCL = - 3 , and U pper C ontrol L imit UCL = + 3 :
Six Sigma 2:15 Average Standard Deviation
+ 3 - 3
Statistical Background (Continued)
At SD = , the number of D efects P er M illion O pportunities ( DPMO ) below the LCL in a normal sample is:
Six Sigma 3:15 In a normal sample, the DPMO will be the same above the UCL. The plot shows f(x) for = 2.
In accepted six-sigma methodology, a worst-case shift of 1.5 in the distribution of quality is assumed, to a new average value of + 1.5
Six Sigma 4:15 In this case, the DPMO above the UCL = 66,807, with only DPMO = 3 below the LCL ( = 2).
However, if is reduced by ½ ( = 1), so that the new LCL = - 6 , and UCL = + 6 , the DPMO for normal and abnormal operation are now much lower: