|
|
To start the process the engineer sets Ref Search (dB) to a non-zero value in the Sequence Maintenance window. Normally the value would be 9.9 dB but if you know the product you may get there a little quicker with a smaller limit. Now, when a unit passes the computer evaluated response test, it is further evaluated. The difference between the test unit and the standard is calculated for every point in the response and then averaged. This figure is then tested to see if it is less than the limit set for the reference search. If it is less then the unit is considered a possible duplicate standard, but first the operator is asked to confirm that they haven't just test the active standard. (An engineer might measure the standard against itself to see if anything has changed.) If the operator confirms that the current test unit is not the standard then they are prompted that the unit is a possible duplicate standard and that they should mark it up and put it aside. Now here's the good bit: the reference search limit is updated to the average difference of the current unit. Testing continues and in time a test unit will appear with an average difference less than the updated limit. If the operator confirms that the unit is not the current standard then they are prompted that this is a possible duplicate standard and they should put it aside and return the previous possible standard to production. The Reference Search limit is again updated. You can see that as testing continues each successive unit identified as a possible standard is a better match than its predecessor is. Typically, in our experience, no further units are identified after the process has been running for some three working days, ie twenty four hours non-stop. |
|