Page 113 - Building A Succesful Board-Test Strategy

Inspection as Test 99

approach has strengths and weaknesses that engineers must consider when decid-
ing on the best strategy.
Unfortunately for test engineers, inspection often falls into the "not invented
here" category, because in many-organizations inspection comes under the purview
of the manufacturing rather than the test department. Any attempt to replace por-
tions of the test strategy with inspection steps is often perceived as a threat to job
security. This attitude emphasizes the necessity of encouraging the "we're all in this
together" philosophy. Buckminster Fuller, engineer and Renaissance thinker, con-
tended that the solution to the problem of trade imbalances around the world was
to simply draw a line around the planet and call it one market. This eliminates the
problem of "them vs. us" because "them is us," so there is no imbalance. The same
philosophy applies to manufacturing and test responsibilities. All quality steps,
including process monitoring and feedback, inspection, test, and repair belong to
the same basic task-—getting only good products out the door. If "them is us," then
including inspection in what would otherwise be "merely" a test strategy creates no
imbalance in responsibility—only a shift in timing. And because inspection usually
precedes test, finding and dealing with problems during inspection costs less than
doing so later.
Lack of access to board nodes often drastically reduces the effectiveness of
bed-of-nails test techniques. Shifting the burden to functional and system test
usually means increasing the time required for test-program development, debug-
ging, and implementation. Once again, pressures to shorten time-to-market strain
the ability to adopt this approach.
Ironically, many common factors both encourage and discourage including
inspection as part of an overall "test" strategy, as Figure 3-2 shows. Smaller parts
and denser board placement make bed-of-nails access more difficult, but they also
complicate the lives of human visual inspectors and require higher-resolution
inspection equipment. The need for higher resolution, in turn, increases camera-
positioning time and image-processing time, which slows the inspection step
and increases the likelihood that lighting or other conditions during inspection
will falsely flag good boards as bad. More complex components mean analyzing
more solder joints placed closer together. Increasing use of EGAs, flip-chips.
and other hidden-node device designs precludes human and automated optical
inspection, because those techniques cannot see nodes out of the line of sight.
The increased use of complex, low-volume boards in today's products favors
inspection because of its shorter program-development time as compared with
conventional test. These same boards require fast ramp-up to full production,
making it unlikely that engineers will have time to fine-tune the inspection
equipment to more easily differentiate between marginally good and marginally
bad boards,
An effective inspection step can verify both a product's quality and reliabil-
ity. As mentioned earlier, quality denotes that a product performs to specification.
Reliability indicates the degree to which the product will continue to perform to
specification—and without failure—during actual use. Test, in contrast to inspec-
tion, concentrates primarily on quality issues.