Approaches to Test Scheduling, Inspection Plans

The inspection plan or Test Frequency can be defined at the variable level and may be overridden by the product specification.

For time-based variables, the sampling interval (minutes) is used as the base interval for scheduling tests. The test frequency in the product specification is used as a multiplier on the base interval. For example, a variable with 60 minute sampling interval and a product specification test frequency of 3 will be scheduled every 3 x 60 = 180 minutes.

For event-based variables the default test frequency is defined at the variable level and is overridden by the test frequency in the product specification. For example a variable with 5 as the default test frequency and a product specification test frequency of 3 will be scheduled every three events.

In addition to base scheduling there are a few other opportunities to modify the base frequency with exception conditions. The schedule for time based variables may be reset at product change, or never reset based on the Reset Test Frequency attribute of a variable. When reset at product change, a test will be scheduled for the interval immediately following the product change and then follow the prescribed frequency after that point. When the frequency never resets, scheduling is always based on midnight as the reference.

Extra testing may be scheduled for tests that fall outside the Warning or Reject limits. Using the Force Test Frequency attribute of a variable, tests may be scheduled immediately after a Warning Test, or a Reject Test, or both. For example, if the test frequency were once every five batches, but a test fell outside the warning limits, a test would be scheduled on subsequent batches until the test fell within the warning limits again.

In determining the testing schedule for event-based variables, the number of good production events into the current production run is used. Any time a production event is rejected, it is essentially ignored in determining the test interval for a given test. If a test were scheduled for a production event that was subsequently rejected, that test would always be scheduled again when the next production event was created.