One of the first quality tools that I learned how to use was the control chart. At the time my job function consisted mostly of receiving a blueprint and designing the process which would ultimately make what was on the blueprint, while at the same time working up an RFP or and RFQ. Being heavily involved in six sigma at the time I frequently went to the gemba and looked at reports from previous production runs and spoke with the employees who ran the machines. One day while speaking with an employee he mentioned that there was a lot of sanding that needed to be done, but that other times yielded no sanding at all. Upon further review I learned that this sanding had been built into the process in order to solve the over-sized parts. Knowing that this was a form of over-processing I decided to collect some data.
As we gathered the data a control chart was used to monitor the process stability as we produced our outputs or the Y's in our process. Eventually we discovered that this was in fact not a common cause variation but rather a special cause of variation. This of course told us that an adjustment needed to be made.
While the design of the mold is not the topic that we will discuss today, the application of the DMAIC method proved to be beneficial. I believe that the key turning point in the life cycle of the part was directly related to using that control chart.
If you have never used a control chart before the control charts general purpose is to monitor if a process stays within a standard over a defined period of time. The other important lesson that can be learned from a control chart is whether or not the process is in control or not, hens the name "control chart." The control chart shows us an upper limit and a lower limit and a mean or an average line. For example if you went bowling and wanted to monitor where the balls were hitting the average would show a cluster of where the ball hit most. As the name hints control charts show us when a process is no longer in control and an adjustment needs to be made. The key is to make the adjustment in relation to the "root cause" and not the issue that is first seen.
In my circumstance I wanted to see the readings and get started on a solution right away which is why the control chart worked so well for me. Most control charts need somewhere between 12 to 25 readings in order to estimate our limits. While every circumstance is different the control chart can often be used to analyze patterns of a process, control a process that is ongoing and to determine whether or not the process should focus on the problem or change the process entirely.
Here are a few basic steps to follow when using a control chart:
1. First define what it is you are trying to solve.
2. Select a control chart.
3. Measure or gather data that will be used for analyzing (If possible place data into control chart in real time).
4. Review data looking for patterns, "out of control points" and get to the root cause.
5. Correct the root cause.
6. Continue monitoring to ensure control.
To set up your own control chart is reasonably simple. Simply follow these steps or you can download a template by signing up below:
Step 1 - Collect data.
Step 2 - Insert data into template.
Step 3 - Calculate mean (center line)
Step 4 - Calculate the UCL or upper control limit.
Step 5 - Calculate the lower control limit.
Step 6 - Select data and insert chart.
In my circumstance of gathering data, the control chart proved to be a valuable tool and helped me to successfully eliminate some inefficiencies in that particular value stream. The control chart can do the same for you too if the situation is right. Don't forget to grab your free template below.