January 16, 2012
Descriptors Gone Wild
Descriptors are powerful tools that are used often when we collect SPC data. They perform many functions including filtering data to a more specific level, assisting in the creation of special reports like the Lot Summary, and keeping track of serialized items. If this were all descriptors did, they would still be one of the most powerful tools available to ProFicient, but descriptors can be used to accomplish much more.
Let's consider a case where I need to perform calculations on my collected data but I have more variables than tests. We could use the factors that are part of the Part, Process, and Test items but that still doesn't give me all the variables we need for our calculation. It may also be that the variables change based on conditions other than Part, Process, and Test so the factor in those items may not always provide the needed leverage. One of the ways we could get around this condition is to use descriptors to, not only to provide a value for calculations, but also to capture more descriptive information about subgroup data.
The data we are trying to collect have at least one test characteristic that is affected by the color of the part, the material the part, is made from or the status of the supplier, or any combination of those items. In a perfect world each combination would have its own part name but there are times where that is not the case. When this happens, look into using descriptors in conjunction with the equation editor in ProFicient to calculate the desired test characteristics. The setup in the data entry configuration would follow the same path of adding descriptors on the Optional Items tab except each descriptor would have a factor value defined to match the value associated with the descriptor choice. In the case above we may make parts that are red, green, or blue and we would assign the colors factors of 2.5, 3.75, and 5.25 respectively.
The same steps could be repeated for both the material type and the supplier status. When added to the descriptor list we would make a note of which item number it is on the list. In the case of ProFicient 4.4 Process is always item 1 on the list followed by Part at item 2. Each additional item on the list is added numerically so for our example Color will be item 3.
On the test list there may be several characteristics that need to be checked during the course of our data entry and for our example we will just use 2 items; one test that we measure with a gage and another test that we calculate using the 1st test result along with the color of the item. Test 1 is set up using a keyboard entry and test 2 will be a calculation. When equation editor opens we see an item in the function list defined as desc(n), meaning it returns the factor of the nth descriptor. We create the equation of test result 1 times the factor of our color descriptor, descriptor 3 in our data entry configuration.
The use of the factor allows us to capture both the descriptive details of the optional item and use the optional item to calculate other results based on the choice. The factor can be up to 16 characters long and can also include exponential values such as 2.5E6 for 2.5 million. So, if you run across a situation where you need a descriptor to do more than just be descriptive, consider using descriptor factor as a means of adding more power to your data collection process.