Page 184 - Six Sigma for electronics design and manufacturing

The Use of Six Sigma with High- and Low-Volume Products and Processes
purpose of these activities is to inform the new product design teams
of the current quality status of different operations in manufacturing
and the supply chain. If the design team finds the process capability
inadequate, manufacturing has to purchase better-quality equipment
or select new suppliers that can meet the quality goals. The process
capability data has to be updated regularly in order to keep design
team abreast of quality and capability enhancements. The frequency
of updates should be short enough to comfortably fit inside the new
product design cycles, as well as meet yearly management goals. A
typical frequency of updating process capability is every quarter. 153
For assembled parts, the process capability determination has to be
compatible with industry standards, as well as the calculations of de-
fect opportunities. For PCBs and their terminations, standards such
as DPMO are used (see Section 4.3.3). For fabricated parts, especially
those made in machine shops, the process capability determination is
more difficult. The machine shop can produce parts with the desired
geometry using many possible machines in the shop; some producing
high-quality parts and others parts of much lower quality. The dilem-
ma is whether a particular process should be machine dependent, es-
pecially since the machine selection is usually not included in the part
or assembly documentation. If a ½ hole needs to be drilled, there are
many alternative machines in the shop to perform this operation,
with varying process capabilities. So what will the design team as-
sume for the ½ holes defect rate?
One solution to the fabrication dilemma is to allow for an additional
attribute in the six sigma methodology. This attribute would be a
quality or complexity indicator. The fabrication shop could be divided
into several (maximum of three) levels of complexity. As each new
part is being designed, the design engineer can select from any of the
three process capabilities available, depending on the level of com-
plexity of the part.
For each process, a baseline process capability is determined, ac-
cording to the sampling methods outlined in Table 5.4. Every quarter,
all of the process capabilities are checked, and recalculated if they
show a statistically significant shift in average or using statistical
comparison tests. The z distribution is used to compare a large (>30)
2
sample with the baseline population averages; the test is used to
compare sample to population . For smaller-size samples, the sample
average shift to the population average can be tested with the t distri-
bution, as shown earlier in this chapter.
Some of the process capability data can be obtained from control
charts, as shown in Chapter 3, whereas others can be calculated di-
rectly by taking samples from the production line. Table 5.5 is an ex-
ample of a production line of PCB assembly process capability calcula-