What Is a PFU? Process Capability Study According to VDI/VDE 2645-3 Explained

Your torque wrench has passed the machine capability study with Cmk = 1.85. The documentation is complete, the measuring equipment is calibrated, the release has been granted. And yet the next customer complaint shows: The bolted joints from series production are outside the tolerance.

This scenario is not an exception - it is the classic result of mixing up two different inspection concepts. Anyone who treats the process capability study (PFU) as identical to the machine capability study (MFU) systematically underestimates all the disturbance variables that occur in real production. This article explains clearly and precisely: What is a PFU, what does the VDI/VDE 2645-3 standard require, and how do you perform a compliant process capability assessment correctly?

Definition: What is a PFU?

VDI/VDE 2645 Part 3 describes methods for process capability studies of residual torque (weiterdrehmoment) on bolted joints with preload. The goal is the assessment and documentation of the quality capability of a tightening process under series production conditions. The PFU provides guidance for evaluating and continuously improving the tightening process.

Specifically, this includes:

• Identifying systematic influences in order to improve the process in a targeted way
• Evaluating the effectiveness of corrective measures
• Defining intervention limits for quality control charts in production monitoring

In contrast to the MFU, the PFU considers not only the machine influence, but also the influence categories man (operator), material, method and environment. This is the well-known 5M model in production control - and in torque control and bolt joint analysis all five dimensions act simultaneously.

MFU vs. PFU: The decisive difference

The most common source of error in assembly quality assurance: MFU and PFU are treated as equivalent or confused. They are not - they answer fundamentally different questions.

The MFU asks: Is the tool fundamentally precise enough?
The PFU asks: Does the complete tightening process reliably stay within tolerance under real conditions over time?

A passed MFU is a basic prerequisite for the PFU - but it does not replace it. The MFU is typically the first step when accepting a new system (Gate 1). The PFU follows later and proves process control in real production (Gate 2). A good MFU value is a prerequisite, but no guarantee for a good PFU value.

Practical example: An electric screwdriver reaches Cmk = 1.85 in the lab. In ongoing series production - with changing operators, temperature differences between early and night shifts and a new screw batch - the PFU only yields Cpk = 0.98. The process is not capable. The tool was - the process never was.

This is exactly why a proper process capability study is essential in real-world assembly testing and torque analysis.

Why the PFU is mandatory for Class A joints

The VDI/VDE 2862 Part 2 standard defines minimum requirements for tightening processes and classifies bolted joints in three categories:

For Class A bolted joints, the PFU according to VDI/VDE 2862 Part 2 is mandatory. This applies to all industries in which bolted joints fulfil safety-relevant functions - explicitly not only automotive, but also aerospace, rail and medical technology.

Missing or invalid PFU evidence for Class A joints can lead to major nonconformities in IATF 16949 audits - and, in the event of damage, to personal liability under product liability law. You can read more about liability in our article Liability risk in screw assembly: Why the "state of the art" is not optional.

The Cpk value: The key metric in the PFU

What Cpk measures

The process capability index Cpk evaluates how far the process mean is from the specification limits - taking into account the actual process variation.

Formula:

Cpk = min [ (USL - x̄) / (3σ), (x̄ - LSL) / (3σ) ]

In this context:

• USL = Upper specification limit (Nm)
• LSL = Lower specification limit (Nm)
• x̄ = Process mean (Nm)
• σ = Standard deviation of the measured values

While the Cp value only expresses the ratio of tolerance width to process variation, the Cpk value additionally considers the position of the mean relative to the centre of the tolerance range. A process can show good repeatability and dispersion but still have a poor Cpk - because its mean is too close to one of the tolerance limits.

Cpk thresholds: What applies in which industry?

A Cpk of 1.67 corresponds to a scrap rate of only 0.57 ppm (parts per million). A Cpk of 1.33 corresponds to 63 ppm.

Important: Cpk applies to the PFU - that is, to the real production process capability assessment. For the MFU, the Cmk value is used. Do not confuse these two indices. You will find more detail on formulas, Cpk calculation and worked examples in our article Cmk and Cpk in tightening technology: How to calculate and interpret capability indices correctly.

Interactive Cpk calculator

Calculate your Cpk value directly for your torque measurement:

PFU in practice: Step-by-step guide

VDI/VDE 2645-3: What the standard specifically requires

VDI/VDE 2645 Part 3 (current version: February 2019, replacing the July 2016 edition) defines the methodological approach for the PFU in tightening technology. The main requirements at a glance:

Measurement method: Residual torque (weiterdrehmoment)

The standard is based on residual torque measurement as the test method. The residual torque is the torque required, after the tightening process is completed, to turn the screw a small additional angle. It gives an indirect indication of the installed preload - and thus of the real joint quality - without destroying the joint. This is a core method in modern bolt joint analysis.

Sample size

For statistically reliable statements, practice recommends at least 125 measured values, taken over a representative production period. A typical error: samples that are too small. Stick to established minimum sizes (e.g. 125 values). Document the sample size and justify it. The confidence interval of the Cpk value becomes very large with small samples - this significantly reduces the reliability of your process capability study.

Process stability as a prerequisite

Only if the control chart over a representative period shows no violations of intervention limits or unnatural patterns (trends, runs) is the process stable and the PFU valid. Evaluating an unstable process does not yield valid Cpk values for assembly quality assurance.

Documentation requirements

The PFU must be fully documented, including:

• Measurement log with all individual values and timestamps
• Control chart as proof of process stability
• Cpk calculation with process mean, standard deviation and specification limits
• Measuring equipment evidence (DAkkS calibration history of the torque tools used)
• Date, operator, production identification for complete traceability

You can find more about documentation requirements for Class A bolted joints here: Documentation requirements for tightening processes: What you have to document.

Common PFU errors - and how to avoid them

In practice, many PFUs fail not because of measurement technology, but because of methodological errors. The seven most common:

1. Confusing MFU and PFU: The Cmk value from the MFU is interpreted as process evidence. It is not - it only proves tool capability under ideal conditions in the lab.

2. Sample too small: 30 or 50 measurement values are generally not enough for a reliable Cpk statement. The confidence interval is too wide.

3. Evaluating an unstable process: If you skip the control chart analysis and go straight to Cpk calculation, you will get a mathematical result - but not one you can trust statistically.

4. Lab conditions instead of series environment: PFU measurements carried out in a quiet test room instead of directly at the production workstation ignore exactly the disturbance variables that the PFU is supposed to make visible.

5. Incorrect specification limits: USL and LSL are estimated from experience rather than taken from the design specification. This systematically distorts the Cpk.

6. Unsuitable measuring equipment: The measuring tool itself has not been assessed for measurement system capability (Cg/Cgk according to VDI/VDE 2645-1). An inaccurate measuring system yields unrealistic Cpk values and undermines reliable torque control.

7. Incomplete documentation: Measurement values exist, but the control chart is missing, the measuring equipment record is outdated, or operator information is not recorded. In an audit, that is not sufficient.

PFU with GWK: The right measurement technology for compliant results

A PFU is only as robust as the measuring system used. GWK provides matched torque tools and systems from a single source - Accuracy by GWK - for professional process capability assessment and production monitoring.

QUANTEC MCS® - The tightening lab for development and QA

The GWK QUANTEC MCS® is specifically designed for test sequences according to VDI/VDE 2645-3 and modern torque analysis:

• Non-fixed-point angle measurement for precise residual torque acquisition
• ±1 % accuracy between 10 and 100 % of the nominal range - fully traceable
• Automatic evaluation and report generation with the QuanLab Pro® software
• Integrated Cpk calculation, histogram and control chart directly in the system
• WLAN data transmission for tamper-proof archiving

QUANTEC MCS® performs both MFU and PFU - and automatically generates complete, standards-compliant documentation for your assembly testing and assembly quality assurance.

OPERATOR® - PFU directly at the production workstation

For PFU under real series conditions, the GWK OPERATOR® is used as a production tool for on-line torque control:

• Testing directly at the workstation - without stopping the assembly line
• Modular square interchangeable drive system for flexible use at different screw locations
• Robust aluminium-titanium construction for continuous production use
• Optional barcode scanner* for part identification and full traceability

*Optional special accessory

Q-CHECK® - QA and audit tool for residual torque measurement

The GWK Q-CHECK® is a QA and audit tool designed specifically for residual torque measurements as part of a process capability study. It is suitable for sample-based checks in production and during audits - with direct printing of standard-compliant test reports.

DAkkS calibration laboratory: The foundation of every valid PFU

GWK operates its own DAkkS calibration laboratory with the fully automatic DWPM-1000® test machine, accuracy class 0.2. DAkkS calibration is the foundation for traceability of your measuring equipment - and thus a prerequisite for an audit-ready PFU. In addition, GWK offers mobile calibration services for on-site calibration in your production - for minimum downtime.

These services ensure that your torque tools and Quantec MCS systems always deliver reliable, traceable measurement results.

Frequently asked questions about the PFU (FAQ)

Conclusion: PFU is not bureaucratic red tape - it is process security

The process capability study according to VDI/VDE 2645-3 is the only instrument that proves whether your tightening process functions reliably under real production conditions. A good Cmk value from the MFU gives you confidence in the tool in the lab. The Cpk value from the PFU gives you confidence in what really counts: the bolted joint on the component - day after day, shift after shift.

Consistent use of PFU lets you detect process drift early, prevent defective parts and protect yourself in audits. Skipping it risks exactly the scenario described at the beginning of this article.

GWK supports you throughout the entire process capability study - from choosing the right measuring equipment, through DAkkS-accredited calibration in our DAkkS calibration laboratory, to standards-compliant documentation with QuanLab Pro®. Together we develop the optimal solution for your specific requirements in torque control and bolt joint analysis.

Further reading:

• MFU vs. PFU: Why machine capability alone says nothing about your tightening process
• Cmk and Cpk in tightening technology: How to calculate and interpret capability indices correctly
• VDI/VDE 2862 explained: What Categories A, B and C mean for your tightening processes
• Audit checklist: 10 points your auditor will check in tightening processes