How to Create a Statistical Process Control Chart

A statistical process control chart is used in manufacturing to evaluate the output or activity of a particular process or work flow and determine if the process is in control or not. In control means current data is an accurate source of future predictions. Out of control indicates that there is an error in the current process than needs correction.

Instructions

1. Identify a manufacturing process or work flow where the product is measured at regular intervals. Good processes to select are those where the output must fall within a predetermine range to be acceptable. Evaluation of the work output against a standard evaluation matrix is required for the results to be of any value. An example would be the power output of 40 watt light bulbs.

2. Calculate the upper control limit, center line and lower control limit. The upper and lower control limits should be the outer limits of acceptable product control. The control line should be in the exact center of the upper and lower control limits, usually within 2 standard deviations from the center line. For the light bulb example, the upper limit is 42 watts and the lower limit is 38 watts. Any values outside these limits is unacceptable. Within this example, the control line is 40 watts.

3. Create the statistical control chart using chart paper or a spreadsheet program. The x axis is used to record the number of samples taken. The y axis shows the upper control, center line and lower control limits, as well as the values in between. Draw the control lines across the entire chart.

4. Collect samples from the production line over a specific time frame. Take this data and map the results from the control test against the control chart and make a line graph. Read the graph and interpret the peaks and spikes. These points indicate the state of the process at that point in time. In the light bulb example, if several samples have readings of 41 watts, the process is moving towards being out of control for the upper control limit.