PPAP Sampling Is Not a Quality System - It's a Starting Point

What PPAP Actually Requires - and What It Doesn't

The Production Part Approval Process, defined by the AIAG PPAP manual (currently 4th edition), requires suppliers to demonstrate that their production process can consistently produce parts meeting customer requirements. The submission typically includes dimensional results from a 30-piece initial production run, a process capability study (Cpk), a failure mode and effects analysis (FMEA), a control plan, and measurement system analysis (MSA) data for key characteristics.

When a customer approves a PPAP submission, they are confirming that, based on the submitted evidence, the process appears capable. They are not certifying that the process will remain capable under production conditions. They are not approving the supplier's ongoing quality monitoring approach. They are approving a snapshot of process capability taken during a controlled pre-production trial.

The gap between PPAP conditions and ongoing production conditions is where quality problems develop. Pre-production trials are typically run with heightened attention from engineering and quality staff, freshly maintained tooling, selected raw material heats, and experienced operators. Regular production involves shift changes, operator rotation, die wear progression, ladle temperature variability, and the hundred other sources of variance that engineering oversight suppresses during the approval run.

The Cpk Submission Problem

PPAP requires process capability studies with minimum Cpk values (typically 1.67 for automotive critical characteristics). The study is conducted on the 30-piece initial sample. The problem is that 30 parts from a single production session is a poor sample for estimating long-term process capability in casting operations where significant variance sources are time-varying - die wear, furnace slag accumulation, operator-specific ladle handling, shift-to-shift temperature variation.

A 30-piece Cpk study can look excellent if those 30 parts are run under favorable conditions. Die temperature is stable at the beginning of a production run. A fresh die has clean cavity surfaces without progressive wear marks. A dedicated operator handles the trial. The Cpk comes out at 1.85. Production approval is granted. Three months later, with worn dies, rotating operators, and normal production variability, the actual Cpk for the same dimension is running at 1.2, below the approved specification, generating customer complaints.

This is not hypothetical. It describes a failure pattern we see documented in supplier corrective action reports (SCARs) across the automotive supply chain. The PPAP approval based on 30-piece capability was not wrong - the process was capable at that moment. What was missing was an ongoing monitoring plan that tracked capability through the full production die life and triggered review before the process drifted out of acceptable range.

What a Real Ongoing Control System Looks Like

The control plan submitted at PPAP is supposed to define ongoing monitoring. In practice, control plans for casting operations often specify inspection frequencies and methods that were adequate for PPAP but are not sufficient for production volumes. A control plan that specifies CMM inspection of critical dimensions at a frequency of "first article + 1 per 500 pieces" is checking 0.2% of production volume. If the process drifts after piece 200, the next CMM check at piece 500 arrives too late to prevent 300 rejects.

For casting operations, an adequate ongoing control system has three components: in-line 100% inspection for surface and dimensional defects detectable at the production rate, attribute SPC on defect rate by type (not just dimensional Cpk), and a trigger-based response plan that specifies what happens when control chart signals appear.

In-line 100% inspection does not mean 100% CMM. CMM cannot run at casting production rates. It means in-line vision inspection for surface defects and dimensional drift monitoring, supplemented by statistical sampling CMM at intervals calibrated to die wear rate. The combination provides high-coverage detection at production speed with periodic precision measurement to catch dimensional drift before it exceeds tolerance.

As discussed in our article on billet temperature variance and defect correlation, SPC on defect rate by type is more informative than aggregate scrap rate. A spike in cold shut rate indicates a different corrective action than a spike in surface crack rate. Aggregate scrap rate charts are lagging indicators of problems that specific defect-type SPC could catch earlier.

The IATF 16949 View

IATF 16949 clause 10.2 requires organizations to react to nonconformities, evaluate the need for corrective actions, and implement those actions in a timely manner. It requires analysis of data from monitoring and measurement activities to demonstrate process effectiveness and identify improvement opportunities.

An organization that approves a PPAP, ships production parts, and only discovers quality issues through customer complaints is not meeting IATF 16949's intent for process monitoring. Customer complaints are the most expensive feedback mechanism - they arrive after defective parts have shipped, been received, possibly been used in assemblies, and triggered warranty or sorting costs at the customer facility.

Third-party auditors increasingly cite inadequate in-process monitoring as a finding in IATF 16949 audits of casting suppliers. The combination of low CMM sampling frequency, no in-line detection, and aggregate-only scrap reporting is a recognized audit finding. The corrective action customers require typically involves adding in-process detection capability - which could have been in place before the complaint that triggered the audit.

First-Pass Yield as the Right Metric

Many casting operations measure scrap rate as the primary quality metric. Scrap rate captures parts that are rejected before shipment. It misses parts that pass internal inspection, ship to the customer, and are rejected at incoming inspection or in service. It also measures only the end result, not the process behavior that produced it.

First-pass yield - the percentage of parts that pass all quality checkpoints on the first pass without rework or reinspection - is a more informative metric. It reflects process consistency, not just end-of-line sorting effectiveness. A foundry with aggressive outgoing inspection can have a low scrap rate and a poor first-pass yield: lots of marginal parts being caught and sorted at the last station before shipping.

Tracking first-pass yield at each inspection point - visual inspection, dimensional check, pressure test where applicable - creates a map of where in the process quality problems develop. This information is essential for directing process improvement effort and for evaluating whether capital investments in tooling, process control, or inspection are producing measurable results.

OEE (Overall Equipment Effectiveness) calculations for casting cells should incorporate first-pass yield as the quality component, not just end-of-line scrap rate. An HPDC cell running at 90% availability and 95% performance but 88% first-pass yield has an OEE of approximately 75% - meaningfully below the 85% benchmark typically targeted in automotive components manufacturing.

The Practical Path Forward

The practical improvement for Tier 2 casting suppliers is not to rebuild the quality system from scratch. It is to identify the gap between PPAP approval conditions and ongoing production conditions, and close the monitoring gap that allows that variance to go undetected.

Concretely: review your control plan for every active PPAP and ask whether the inspection frequency specified would detect a process shift within 50 parts. If not, the monitoring frequency is insufficient for the production volume and defect risk. For high-volume HPDC operations producing automotive structural parts, 100% in-line inspection is the only way to meet this standard at production rates.

Contact hi@forgepulsx.com if you want to discuss what in-line inspection coverage looks like for your specific casting operations and customer quality requirements.