In addition to making sure that testability receives equal consideration, the IPD team needs to determine several items that will contribute to an effective testability design. For instance, the need exists to define what constitutes a failure. In particular, failures that can affect BIT performance, such as drift, must be clearly defined. This has been a problem that has plagued BIT performance in the field. BIT algorithms that are too sensitive may detect and report failures that only occur intermittently due to environmental or other factors, but it may not be possible to duplicate in the maintenance environment the conditions that caused the failure. A formal process must also be in place to ensure that test verticality is maintained from one maintenance level to the next.
Another area that needs clear definition is in which failures need to be reported by BIT. Should all BIT failures be reported, or only those that degrade safety or mission capability? For example, should a failure be reported in a connector if it occurs intermittently (e.g., an average of once every three flights) or only if the connector has failed three times in five consecutive test attempts? All reportable BIT indications should be carefully reviewed to define the failure state and the appropriate action.
Testability needs also should be determined from field and manufacturing data on like systems. This is a problem, however, as many data collection systems do not adequately report testability problems. Therefore, the data collection system must be devised to collect data such as CNDs. These data must then be analyzed to determine the root causes of such behavior such that corrective actions can be implemented in next generation designs.