close window

Process variation

1. Process variation has two basic causes:

Common causes are variations due to operating environment (temperature, humidity and atmospheric pressure), small equipment vibrations, small variations in the materials used and so on. When a process varies in such a way, over time these variations become predictable. In general, as long as the outputs of a process lie within the expected amount of variation (from common causes) then the process is said to be in control.

Special causes are abnormal and cannot be predicted. In a manufacturing process special causes are excessive machine or tool wear and drifting from calibration, an inferior batch of raw materials, a poorly trained operator or an incorrect work method. Any of these cause variation in the process output and unless rectified could harm the output quality of the process. Whenever a process is judged to have been influenced by a special cause, it is said to be out of control.

2. The two types are Basic and X-bar and R charts

Basic control chart

This is used to show variations of sample data (measurements, temperature etc.) against time or sample number. It would be used if the process is well established and statistical information is known about the process.

X-bar and R charts

If a process is not well established, then X-bar charts will show how the average of a process sample data varies. To complement the X-bar chart (they are usually used together) an R chart is a plot of the variation in the sample data. Any drift in the averages (the averages rise or fall) or excessive variation between samples would prompt investigation.

3. A plot of the figures given is shown below. At first the process is within the 246–250°C limits but at around samples 10 and 11 the temperature starts to drift out of control. Subsequent measurements show no sign of the process coming back into control. In this case the process should be stopped and further investigation to take place.

[ back to top ]