PPAP for sheet metal parts: scope, levels and evidence

An eighteen-element PPAP checklist arrives with a drawing for a 3 mm bracket at 4,000 a month. The part is a bend, four holes and a weld nut. One element belongs to the customer's design analysis, four more apply only if there is a change, a special characteristic or a visible face, and the rest still need evidence that a real inspection can produce.

Production part approval is a real process with real evidence behind it, and it is routinely applied to parts where half the file is filler. The useful skill is knowing which six elements carry the risk on a fabricated part, and what the numbers in them actually demand.

The standards and compliance hub covers first article inspection and material certification, which are the two documents a PPAP submission is built on.

What PPAP is, and why the count of elements matters

The Production Part Approval Process is an AIAG manual, now in its fourth edition, that defines how a supplier demonstrates to a customer that a production process can make the part to the drawing at the quoted rate. The output is a part submission warrant, supported by evidence. IATF 16949 makes product approval a requirement for automotive suppliers, which is how an 18-element checklist ends up attached to a bracket.

The manual lists eighteen elements. The number is not a target; it is an index, and several entries only exist because the automotive model assumes the supplier owns the design or subcontracts special processes.

Two scoping points decide most of the argument. First, PPAP is a customer requirement, not a certification, so the customer's own supplement can add or drop elements. Second, it approves a process at a point in time, not a company.

The eighteen elements, and which ones a fabricated part actually needs

The right-hand column is the practical filter. "Customer" means the evidence belongs to the design owner and the supplier only references it.

Read the table downwards and the shape of a real submission appears: a process flow that matches the quoted route, a process FMEA that names the actual failure modes, a control plan that puts a frequency and a reaction plan against each characteristic, and a capability study on whatever the customer calls special.

The three numbers that decide acceptance

Capability indices are where PPAP submissions get rejected, and the arithmetic is simple enough to check before the run.

Ppk of 1.67 means the tolerance band spans ten standard deviations, because 6 × 1.67 is approximately 10. On a ±0.5 mm tolerance the band is 1.0 mm, so the standard deviation on the submitted run has to be 0.100 mm or less. A process running at σ = 0.12 mm reaches Ppk 1.39 and fails, even though every part in the sample is inside tolerance.

Cpk of 1.33 is the usual ongoing requirement, and it spans eight standard deviations. On the same 1.0 mm band, σ can rise to 0.125 mm.

The distinction between the two indices is what makes them useful on sheet metal. Ppk uses the overall standard deviation, which includes drift between coils, between shifts and between lots. Cpk uses the within-subgroup standard deviation, which sees only the variation inside a short run. A progressive die running one coil can show Cpk 1.8 and Ppk 1.3 across three coils, because coil-to-coil thickness and hardness move the dimension in a way that a five-piece subgroup cannot see. When a customer asks for Ppk, they are asking about the whole 300-part run, not the best hour of it.

The run itself is defined: production tooling, production process, production rate, and a minimum of 300 consecutive parts unless the customer agrees otherwise. Three hundred parts gives 60 subgroups of five, which is enough to calculate both indices and to plot a control chart that means something.

Submission levels 1 to 5

The level decides how much evidence the customer receives, not how much the supplier produces.

• Level 1 — the warrant alone, with the supplier holding the supporting data.
• Level 2 — the warrant, sample parts and limited supporting data.
• Level 3 — the warrant, sample parts and complete supporting data. This is the default in the manual and the level most programmes use.
• Level 4 — the warrant and whatever else the customer specifies.
• Level 5 — the warrant, samples and complete data, reviewed at the supplier's site.

Level 1 is the cheapest submission and the weakest evidence, because the customer sees a signature and nothing behind it until an audit. Level 5 is the most expensive and is usually reserved for a new plant, a new process or a part with a history of problems.

The level is also not permanent. A design change, a process change, a tool transfer or refurbishment, a new sub-tier source, a move to another site or a production gap of more than twelve months triggers a resubmission, and the trigger list belongs in the purchase order rather than in a supplier's memory.

The control plan is the element the shop actually runs

Of the eighteen elements, the control plan is the only one that governs the part after the approval meeting. It is also the one most often copied from a previous submission and left unedited.

A usable control plan carries five fields per characteristic: the specification, the measurement method with the gauge, the sample size and frequency, the control method, and the reaction plan.

For a 3 mm bracket with three special characteristics — two hole positions and a weld nut height — a workable plan reads: check the hole positions on five parts at the start of each lot and every 500 parts, in the checking fixture; plot the values against control limits; if any point falls outside, stop the lot and quarantine it back to the last good check; if two consecutive points sit more than one standard deviation on the same side of the mean, re-qualify the tool before running on.

The frequency arithmetic is what keeps a plan alive. Five parts per lot at three minutes each is 15 minutes of inspection. On a 1,000-piece lot that is 1.5% of the labour content, while a 100% check on the same part would consume more than 50 hours. A plan that the shop cannot afford gets signed and then ignored, and an ignored plan is worse than no plan because the file states that a check happens.

Where PPAP and a first article inspection differ

Both approve a part, and they are not interchangeable. A first article inspection to AS9102 records every characteristic on the drawing on a small number of parts from the first production run, and it is a conformance record. A PPAP approves the process with a control plan, a failure mode analysis and a capability study, and it is a production readiness record.

The dimensional element inside a PPAP is often a full layout on the submission parts, which looks like an FAI. The difference is what comes after it. PPAP asks how the process will behave over 300 parts and what happens when it drifts; FAI asks what the first parts measured.

Where a programme needs both, run the layout once and use it for both documents. Duplicating the measurement wastes the inspection budget and introduces a second set of numbers to reconcile.

What an industrial buyer should ask for instead

For a part that is not going into an automotive programme, a full PPAP is usually the wrong instrument, and a shorter pack gives more useful evidence per rupee:

1. A full first article on three parts from production tooling, with every characteristic measured.
2. A process flow diagram that matches the quoted route.
3. A control plan naming the characteristics that matter at assembly, with a frequency and a reaction plan.
4. Material and finish certificates tied to the lot by heat number and coating batch.
5. A capability study on two or three key characteristics over 100–300 parts.
6. A signed declaration that the submitted parts were made with the production tool, at the production site, at the quoted rate.

That pack is proportionate for an enclosure, a canopy frame or a rack, and it answers the questions a PPAP is really asking: can this supplier hold the part, and what happens when it does not.

What Ganesh runs

Ganesh produces brackets, frames, panels and enclosures from 0.8 mm to 40 mm on presses from 20 T to 300 T, a 250 T punching line, a 12 kW fibre laser with a 40 ft × 16 ft bed and a 200 T, 4,450 mm brake to 10 mm. Tooling, welding fixtures and checking fixtures come from the in-house tool room, so the process on the flow diagram is the process in the building. Three CLOOS robotic MIG cells and more than 30 manual stations cover the welded assemblies, and seven-tank pretreatment with an 18 ft × 14 ft powder booth covers the coated ones. The stamping capability sets out the press and tool room scope.

PPAP and first article inspection are available, and the works hold ISO 9001:2015 and ISO 14001:2015. Automotive programmes run through the same route, with Mahindra and Minda among the customers whose parts carry this level of documentation.

The boundary is worth stating. Ganesh does not own the design on a build-to-print part, so the design FMEA and the customer's engineering approvals come from the customer. Capability studies are run on nominated characteristics; a request to study all sixty characteristics over 300 parts multiplies the inspection cost by an order of magnitude and is almost never what the drawing intended.

What to put on the drawing and in the approval request

• The special characteristics, marked, with the capability requirement for each.
• The tolerance and the datum scheme, so the study measures the same thing the assembly uses.
• The submission level and the customer supplement that applies.
• The sample size and whether parts from each die cavity are needed.
• The material certificate type, and the coating acceptance criteria.
• The change triggers that require a resubmission.
• The retention period for the master sample and the records.

The stamping tolerance guide covers how those characteristics are set, and the inspection article for repeat production covers the sampling that follows approval.

When PPAP is the wrong tool

At 200 parts a year, a 300-part capability run consumes a year and a half of demand, and the study describes a process that will have changed by the time the volume arrives. Low-volume work is better approved with a first article and a control plan.

For a prototype or a one-off fixture, there is no process to approve. The right document is a dimensional report and a material certificate.

And where the customer's checklist has been copied from an automotive programme onto a part with no special characteristics, the honest move is to ask which elements they actually need. Five of the eighteen are customer-owned or conditional, and a supplier who ticks them anyway is telling you something about how the rest of the file was assembled.

Send the drawing, the annual volume and the checklist on a quote request, and the engineering team will mark the elements that apply before quoting the run.