Page 130 - Building A Succesful Board-Test Strategy

116 BUILDING A SUCCESSFUL BOARD-TEST STRATEGY

Once the board is short-free, the tester applies power and an input/output
stimulation pattern. Testing a cellular phone board, for example, would verify
dialing, "press-to-talk," and other features. Again, any faults should exhibit
obvious differences from the good-board thermal signature.

3,3.4.4 No Good Deed . . .
Like other techniques, infrared inspection has its drawbacks. Chief among
them is the need for up to 30 production boards from which the system assembles
the good-board signature. Many manufacturers never see that many good boards
until after production begins, by which time the test must already be in place.
Regardless of the cost benefit of anomaly detection, a manufacturer may encounter
the "chicken-and-egg" problem—needing a test to generate 30 good boards and
needing the 30 good boards to create the test. In high-mix, low-volume, and low-
cost situations this requirement could prove prohibitive.
An out-of-tolerance resistor or capacitor will not generally produce a thermal
signature sufficiently different from a good one to be detectable. Such failures are
relatively rare, however, and an in-circuit test will usually identify them prior to
infrared inspection.
An infrared detector cannot see through certain barriers, such as RF shields
and heat sinks. Such boards would require testing before attaching these parts,
which may be impractical, and will certainly miss any faults induced during
attachment.
Infrared inspection can identify the component containing a failure, but its
resolution is not always sufficient to identify the exact pin location, especially for
solder problems on small-pitch surface-mount boards. In addition, the thermal
anomaly might not occur exactly at the fault. However, a repair technician can call
up the failing thermal image on a computer monitor to examine it before pro-
ceeding. The anomaly's location narrows the search for the actual fault to a few
pins. With that information, the technician can pinpoint the actual problem.
Since the technique depends on such small changes in temperature, the board
under test must be thermally stable. That is, before power-up, the entire board must
be at equilibrium at the room's ambient temperature. A board fresh from wave or
reflow solder, for example, or from storage in a room whose ambient temperature
is more than ±5°C different from ambient on the test floor, must be allowed to
reach equilibrium before the test can proceed. As a result, an infrared detection
system may work best in a batch rather than an inline production configuration.
A preconditioning chamber where up to 45 minutes of production can reach
thermal equilibrium prior to inspection can alleviate this problem. The chamber,
however, adds time to the production process and introduces another handling
step. Also, current infrared solutions require a human operator to load and
unload the boards, precluding their use in unattended high-speed automated pro-
duction lines.
The application of infrared technology is new to inline loaded-board inspec-
tion. Early returns are encouraging, and this alternative deserves consideration. Its