Variation that is built into any system is called a "stable system of chance causes" or "common variation." Because this variation is built into the system, it is always present. The only way to reduce the amount of common variation is to redesign and rework the manufacturing system (this usually requires massive effort in management commitment or new technology intervention). A partial list of some nonassignable causes associated with common variation includes things that are:
Normal
Steady
Natural
Predictable
Homogeneous
Same
Not changing
Consistent
Nonsignificant
Statistically constant
Stable
A process that operates with only common variation is systematic and is designated as being in control. A series of samples collected from an in-control process would estimate the same population and would be similar to Figure 6.2.
The curves that represent an in-control process are:
Repeatable
Stable
Consistent
Predictable
The only requirement for process control is that the process performs the same time after time. A process is neither good nor bad just because it is in control; it is merely repeating itself. For example, a process that always makes 20% defective (out of specifications) parts , regardless of material supplier, operator, machine, setup, shift, or environmental changes, is in control (see Figure 6.3).
Although this is not desirable, it fits the definition of an in-control process. The operating level of this process must be changed (or made out of control) if the process is to be improved. Because the problem is part of common variation, a change of the manufacturing system or design is required. Simple adjustments by an operator to one of the process components will not cause a lasting improvement. The sources of this problem are built into the system itself, so the solution must be designed into the operation.