Statistical Process Control (SPC) is a way of watching a process with data so you can see when it changes, and react before it makes bad parts. Process capability, usually reported as Cpk, tells you how well a stable process fits inside the tolerance.
Both are core tools in automotive quality and are widely used in aerospace and general manufacturing. This guide covers the basics you need to set them up and read them correctly.
What is SPC?
Every process varies. Some variation is normal, the result of many small causes that are always present. This is called common cause variation. Other variation comes from a specific change, such as a worn tool, a new batch of material or a wrong setting. This is called special cause variation.
SPC uses control charts to tell the two apart. When the chart shows only common cause variation, the process is stable and the operator should leave it alone. When it signals a special cause, the operator should find and fix the cause.
Choosing a control chart
The right chart depends on the type of data and how it is collected.
| Chart | Data type | Use when |
|---|---|---|
| X-bar and R | Variable (measured) | Small subgroups, typically 2 to 9 parts at a time |
| X-bar and S | Variable (measured) | Larger subgroups, typically 10 or more parts |
| Individuals and moving range (I-MR) | Variable (measured) | One measurement at a time, such as batch processes or slow cycles |
| p and np | Attribute (pass or fail) | Proportion or count of defective parts |
| c and u | Attribute (count of defects) | Number of defects per part or per unit |
Control limits are not specification limits
Control limits are calculated from the process data, usually at three standard deviations either side of the average. They describe what the process actually does.
Specification limits come from the drawing. They describe what the customer needs. A process can be in control and still make parts out of specification, or be out of control while every part is still in tolerance. Never draw the specification limits as control limits.
Once the limits are calculated from a stable period, approve and lock them. Recalculating every time new data comes in hides the very changes SPC is meant to catch.
Run rules: when to react
A point outside the control limits is the most obvious signal, but not the only one. Common run rules, often called the Western Electric or Nelson rules, include:
- One point beyond three sigma.
- Two out of three points beyond two sigma on the same side.
- Four out of five points beyond one sigma on the same side.
- Eight or more points in a row on one side of the centre line.
- Six points in a row steadily increasing or decreasing.
Each signal should lead to a recorded reaction: what was found, what was done and whether the parts made since the last good check need to be contained.
Cp, Cpk, Pp and Ppk explained
Capability indices compare the spread of the process with the width of the tolerance.
- Cp compares the tolerance width with the process spread (six standard deviations). It ignores where the process is centred.
- Cpk takes the centring into account. It is the smaller of (upper limit minus mean) and (mean minus lower limit), divided by three standard deviations.
- Pp and Ppk use the same formulas with the overall standard deviation of all the data, instead of the short-term variation within subgroups.
Cp and Cpk show what the process could do in the short term. Pp and Ppk show what it actually delivered over the study period. A large gap between Cpk and Ppk usually means the process shifts between subgroups.
Capability only means something for a stable process. Check the control chart first: if it is out of control, the capability number does not predict the future.
Typical capability targets
| Index value | What it usually means |
|---|---|
| Below 1.00 | The process makes parts outside the tolerance. Containment and improvement are needed. |
| 1.00 to 1.33 | Marginal. The process just fits and needs close control. |
| 1.33 or above | A common target for ongoing production. |
| 1.67 or above | A common target for initial studies and special characteristics, for example in PPAP. |
These are common automotive targets. Your customer's requirements always take priority.
Making SPC work on the shop floor
- Chart the characteristics that matter: special characteristics and those linked to known failures.
- Make sure the gauge has passed MSA before trusting the chart.
- Let operators enter results at the machine and see signals straight away.
- Define the reaction plan for each signal in the control plan.
- Review capability regularly and act when it drops.
OperaHive SPC software puts control charts at the machine, applies run rules as data is entered, uses approved control limits and calculates Cp, Cpk, Pp and Ppk for every characteristic.