
An inaugural winner of the united states malcolm baldridge national quality
award.Motorola began
its persuit of sixsigma quality standards in everything they didboth
manufacturing and nonmanu
facturing operations.They coined the phrase from the greek character sigma;
that statistices use to
represent units of standard deviation. Statistical tables define how much
area is under various types
of distribution curves at any distance from the mean at function of sigma.
table 8.3 option yielding process capability index=1.0

base case 
a  b  c 
 c_{pk}=0.5 
increase specification 
reduce process average 
reduce process variation 
process average 
9.5  9.5  5.3  9.5 
standard deviation  2.9  2.9  2.9  1.5 
uppre control limit  18.2  18.2  14.0  14.5 
specification  14.0  18.2  14.0  14.0 
Motorola adopted the most common curve, the normal distribution, as a model to define
the probab ility of observing defects. It based its terminology on the performance
of a process with its average shifted 1.5 σ from the center between the specification
limits. Its sixsigma goal translate into merely 3.4 defects per million
opportunities(3.4ppm),which its equivalent to 99.9997 percent per fect. Motorola's
elegant approach enables assigning a quantitative value to process performance
without demanding a strict statistical analysis.
Progress towards the sixsigma goal can be measured by assigning an equivalent
performance level(number of sigmas) based on the measured defect rate.
For example, suppose a process is producing 62
defects per 10,000 opportunities(6,200 ppm defects).This is equivalent to
99.4 percent perfect.
The sixsigma type of analysis is a simplified version of the previously
described process capability index . Rather than requiring the statistical
calculation of process variation, the sixsigma type of performance level
can simply be looked up in a table or on a graph. Furthermore, the sixsigma
type of analysis can be applied to products and services that lack physical
or chemical tolerances.
<< >>

