My company is using ProFicient to collect SPC data and we are learning more about our processes every day. So far my data collection process has focused on variables type data as that is what my customers, internal and external, said they needed to track the process performance. As it turns out, though, we also need to collect data on other types of data, specifically attribute data, that we have via our visual inspection processes.
When we perform our visual inspections there are three levels of visual defects that we look for: those that have been identified as critical, others that fall into a major category, and some that are deemed as minor. We have interest in all three types and want to separate them by type as we collect the data. Additionally, there are different limits attached to each type of defect and we want ProFicient to alarm and send emails when these various limits are exceeded.
We collect our visual data at different intervals than we collect our variables type data. At a minimum we should have a separate data entry configuration for this data. We could take a simple approach and treat each defect type, i.e. critical, major, and minor, as a separate data entry configuration in the project. This would give us an attribute chart for each type of defect and we could create Pareto charts to further assist in our data analysis. While this would certainly be one effective way to collect this data, let's look at some options that are available in ProFicient to not only collect this data in a single data entry configuration but to allow the system to alarm for each defect type separately.
For one exercise we will treat our data as defectives type data with a subgroup size of 50 and the following requirements: ProFicient will alarm if we get a single critical type defective, 5 major type defectives, or 8 minor defectives. As we will be collecting defectives, we will be utilizing either a P chart or an NP chart. This part of the exercise will consist of creating a data entry configuration with a subgroup size of 50 and the three types of visual results -- Critical, Major, and Minor -- listed in any order on the test list and with each test pointing to its own defective list. On the Advanced Options tab, Option B1 - Combine Multiple Code Lists, will be enabled to allow all the different codes to be viewed at once.
The last step in the data collection setup is to define when alarms will happen. We mentioned earlier that we want an alarm if we get a single critical type defective, 5 major type defectives, or 8 minor defectives. In this case, the specification limit is the total for the subgroup so we will start by using the Special Limits tab of the specification limit record. We also need to convert the maximum number of allowable defectives to a percent defective number for the subgroup. This conversion will give us an Upper Subgroup limit of 0.15 for minor defectives, a value of 0.09 for major defectives, and a value of 0 for critical defectives. As each defective item in a subgroup of 50 items effectively means each represents a 2% rate, 8 defective items in a subgroup of 50 equals 16% defective in the subgroup or 0.16 so we will use 0.15 which will cause an alarm at 0.16 or 8 items. Using the special subgroup limit allows ProFicient to add all the same types of defects together and only alarm on the total for the subgroup.
When a user enters the results into the project they will input the number of defective items found and when the defective code list opens they will get to choose from all three defective groups.
Each code will appear on the list with the group name in parenthesis. The user will enter the number of each type of defective they find as a result of their visual inspection. If the total exceeds the percent defective listed for each type -- 8 for minor defectives, 5 for major defectives, or a single critical defective -- ProFicient will create an alarm and an event for the subgroup.