Even in a stable molding process, no two parts are exactly the same. Small differences in temperature, pressure, and material behavior create variation from shot to shot.
That variation isn’t random chaos. It follows predictable patterns that can be measured and controlled.
Standard deviation (σ) measures how spread out your data is around the average.
A small σ means parts are tightly grouped near the mean, indicating a consistent process. A larger σ means wider variation, which increases the risk of parts moving out of tolerance.
This gives engineers a direct way to quantify how stable the process really is.
When measurements are plotted, they typically form a bell-shaped curve known as a normal distribution.
Most parts fall near the average, with fewer parts appearing as you move farther away. This pattern is consistent across well-controlled molding processes.
The shape of that curve reflects process behavior. A narrow, tall curve indicates tight control, while a wide curve signals more variation.
In a normal distribution, about 99.7% of all data falls within three standard deviations above and below the mean.
This creates a total spread of six standard deviations, often referred to as 6σ.
In practice, this defines the natural limits of the process. Almost all parts will fall within this range unless something changes in the process.
Process capability metrics like Cp and Cpk are built directly on standard deviation.
The tolerance range is compared to the process spread (6σ). If the tolerance is much wider than the variation, the process can consistently produce in-spec parts.
If the variation approaches or exceeds the tolerance, defects become more likely.
If a part dimension has a mean of 2.01 mm and a standard deviation of 0.005 mm, the natural variation range (±3σ) is ±0.015 mm.
If the allowed tolerance is ±0.05 mm, the process variation is much smaller than the tolerance. You end up with a process that consistently produces acceptable parts with margin to spare.
If the mean shifts, capability drops even though σ stays the same. This is how variation and centering work together.
Changes in σ often point to underlying process issues.
Increases in variation can come from inconsistent material drying, uneven cooling, machine instability, or fluctuations in melt temperature. These changes may appear before parts go out of tolerance.
Tracking σ helps detect problems early, before they turn into defects.
Understanding the normal distribution allows engineers to define control limits and monitor performance.
By knowing where most data should fall, it becomes easier to spot when the process begins to drift or behave unexpectedly.
This forms the foundation for statistical process control and ongoing monitoring.
Standard deviation is tracked alongside dimensional data throughout development and production.
During validation, it confirms that the process is stable. During production, it acts as an early indicator of change.
That continuous visibility keeps variation under control and supports consistent part quality over time.