Page 135 - Six Sigma for electronics design and manufacturing

Six Sigma for Electronics Design and Manufacturing
104
There is a need to translate the DPU of each operation into a DPU
for the next level.
For product design and manufacturing process engineers, it is
much more useful to communicate and plan manufacturing tests
using process and product yields as opposed to DPUs for the higher
levels of product.
There is a need to manage the conversion of DPUs into yields.
To address these issues, particular industries have developed the
concept of defects per million opportunities (DPMO). These are stan-
dards that define the total defect opportunities per particular product
or assembly. They use specific methods to combine the DPUs of parts
and manufacturing operations, to arrive at the total number of oppor-
tunities. Opportunities can be defined in terms such as:
Opportunities are characteristics or features of the product or the
manufacturing process.
Opportunities must be measurable and have a standard or specifi-
cation with which they can be compared.
Opportunities must be appraised. If a product has features that are
not appraised, they should not be counted as opportunities.
Opportunities are assumed to be independent.
There cannot be more defects in a unit than opportunities.
The opportunity count for a product is constant until the design or
the manufacturing process changes.
An example of a DPMO methodology is the Institute of Printed Cir-
cuits (IPC) Standard 7912 for calculations of DPMO for PCB assem-
blies, which will be discussed later in Section 4.3.

4.2 Determining Manufacturing Yield on a Single
Operation or a Part with Multiple Similar Operations
The manufacturing yield determination is based on the definition of
the probability of obtaining a defect. The FTY is the percentage num-
ber of units produced without defects, prior to test or inspection. It is
different than the traditional yield, which includes rework and repair.
The Poisson distribution, as discussed in the previous chapter, is a
good basis for calculations of defects, especially when the number of
possibilities or outcomes of defects is large and the probability of get-
ting a defect at any time or region is small. In this case, the Poisson
distribution can be simplified from Equation 3.7 as follows: