SPC — Statistical Process Control

SPC uses statistical methods — primarily control charts — to monitor manufacturing processes in real time, detect variation, and distinguish between natural process variation (common cause) and unusual events (special cause). It enables data-driven decisions about when to intervene in a process.

Why companies use it

·Reduces scrap and rework by catching process drift before it produces out-of-specification parts
·Required by IATF 16949 and customer-specific requirements from Ford, GM, and Stellantis
·Provides objective evidence that a process is stable and capable (Cp, Cpk metrics)
·Enables continuous improvement by quantifying process variation over time

What hiring managers look for

·SPC knowledge separates engineers who react to defects from those who prevent them
·The ability to select the right control chart (Xbar-R, Xbar-S, IMR, p-chart, c-chart) shows statistical competence
·Interpreting Western Electric rules to identify special cause variation is a key skill
·Understanding process capability indices (Cp, Cpk, Pp, Ppk) and their limitations is expected

Typical interview questions

What is the difference between Cp and Cpk, and why does it matter in practice?

When would you use an Individuals and Moving Range (IMR) chart versus an Xbar-R chart?

Your control chart shows 8 consecutive points on the same side of the centreline. What does this indicate and what do you do?

How do you set control limits? Why should you never use specification limits as control limits?

Describe how you have used SPC to drive a process improvement. What did you measure and what changed?

Common mistakes

·Confusing control limits (process voice) with specification limits (customer voice) — these must never be mixed
·Using SPC on a process that is not yet stable — control charts on an unstable process give meaningless results
·Setting control limits using too few data points (minimum 25 subgroups needed for reliable limits)
·Ignoring non-random patterns on control charts and only reacting to points outside control limits
·Calculating Cpk without verifying that the data is normally distributed

Real engineering example

A semiconductor packaging line was experiencing 2.3% yield loss from die-attach voiding. By implementing SPC on epoxy dispense volume and cure temperature, the team identified that a special cause (a worn dispenser needle causing volume drift) was responsible for 70% of the defects. Replacing the needle on a scheduled basis, informed by SPC control limits, cut voiding to 0.4% within eight weeks.

Topics covered

SPCcontrol chartCpkCpstatisticalqualityIATF 16949