Page 148 - Six Sigma for electronics design and manufacturing

Determining the Manufacturing Yield and Test Strategy
117
When using implied Cpk, it is assumed that defects are occurring
because of violation of a particular or a composite set of specifications.
The composite specification can be one-sided or two-sided, depending
on the interpretation of the defects. For example, a wire bond could be
treated as one-sided, since it is assumed that in testing the bond, only
a minimum specification value is given. For solder defects, a compos-
ite specification can be assumed to be two-sided, since solder defects
can be caused by too much solder (solder shorts), or too little solder
(insufficient solder defects). The difference between implied one- or
two-sided specifications is that the number of defects representing the
f(z) value under the normal curve should be halved for two-sided spec-
ifications or used directly for one-sided specifications, resulting in dif-
ferent implied Cpk interpretations. The decision for one- or two-sided
specifications for implied Cpk should be left to the appropriate design
and manufacturing engineers. A description of the use of such an im-
plied Cpk process is given in Chapter 2.
The use of an implied Cpk process in assembly line activities is sim-
ilar to the DPMO process. Individual manufacturing processes are an-
alyzed for quality, with a DPU (PPM) and an implied Cpk calculated
for each. For each assembly, such as a PCB running through the line,
the parts counts and process steps are calculated, then multiplied by
the DPU rate to obtain the defects for each step in the assembly line.
Finally the defects are added and then reflected in a total yield using
Equation 4.8 and an implied Cpk. Alternately, the yields could be cal-
culated for each step, then multiplied together to form a total yield. A
decision has to be made for each process as to the type of quality data
to be collected. In the PCB assembly line case, the choices of the qual-
ity data for each process can be as follows:
The use of a particular defect parameter for each process step. For
component types, defect data can be collected on the following: axi-
al insertion for through-hole components, pick and place operations
for SMT components, odd-shape components for automatic as well
as manual placement, and mechanical parts assembly such as with
screws and special connectors. For terminations, defect data can be
collected on manual as well as automatic soldering.
This division of defect data according to the process used can help
in identifying lower-quality process steps and in targeting these
processes for quality improvements.
Data collection can be based on a selected attribute. For example,
placement quality data can be based on components, leads, or a
combination of both.
Guidelines have to be established in order to handle defects from
prior operations that might influence defects in subsequent opera-