SPC (Statistical Process Control)

Statistical Process Control. A method for monitoring, controlling, and improving a process through statistical analysis. The purpose is to identify and eliminate special causes of variation to ensure stable and predictable production

Cpk (Process Capability Index)

A measure of how well a process can produce parts within the tolerance limits when the process is stable and under statistical control. Cpk accounts for both the process variation and its centering relative to the tolerance limits

Ppk (Process Performance Index)

“Process Performance Index. It measures the actual performance of a process over time, regardless of whether it is under statistical control or not. Unlike Cpk, which measures inherent capability, Ppk accounts for all variation (Total Sigma) that has actually occurred during the measurement period

Cmk (Machine Capability Index)

Machine Capability Index. It is used to verify a machine’s technical ability over a short period of time and under ideal conditions (often by measuring 50 consecutive parts). It shows the machine’s inherent precision without the influence of external process factors.

Sigma (σ)

Standard deviation. A statistical measure of the spread within a data set. In SPC and Six Sigma, the goal is to reduce the standard deviation so that the process variation fits well within the customer’s tolerance requirements

Normalfördelning

A bell‑shaped distribution (Gaussian curve) where most values cluster around the mean, and the frequency decreases symmetrically the further you move from the mean. This forms the basis for most capability calculations in industry.

Skillnad USA vs Europa (Ppk)

There is often some confusion surrounding the concept of Ppk. In the US (according to AIAG/IATF), the term Ppk is commonly used to describe actual performance based on Total Sigma for the entire data set. In Europe (according to VDA/ISO), the definition is essentially the same, but ISO standards are often more strict in stating that Ppk should primarily be used for processes that have not yet been proven to be in statistical control (for example, during a pre‑production run). The biggest practical difference usually lies in which statistical estimate (Sigma) is used in the calculation.

Potential (Cp/Pp) vs. Actual Performance (Cpk/Ppk)

Indexes without the ‘k’ (Cp, Pp, Cm) measure the process’s… potential “— that is, how good it could be if it were perfectly centered within the tolerance. Indexes with the ‘k’ (Cpk, Ppk, Cmk) measure the actual performance…” utfallet …by taking into account how close the process actually is to the worst‑case tolerance limit.