Welded 316L Sensor Tube: PPAP Level 3 Case

What this part is

A thin-wall stainless tube closes at one end, gets a welded joint, is ground to a controlled outside diameter, trimmed to final length, and vibratory-polished. Parts of this shape sit in temperature and pressure sensor assemblies, where the tube both protects the element and — because it is welded shut — has to hold pressure. That is why the drawing carries both a hardness callout and an 8 bar leak requirement: the part has to be formable, weldable, and airtight at the same time.

Drawing requirements

Characteristic	Requirement	Control method
Material	SUS316L	Mill certificate per lot, EN 10204 type check
Overall length	106 ± 0.5 mm	Caliper, 5 pcs per check
Outside diameter (ground)	Ø4 +0 / −0.05 mm	Caliper, 5 pcs per check
Inside diameter	Ø3 ± 0.05 mm	Go / no-go plug gauge
Hardness	HV 300 ± 50	Hardness tester, 5 readings
Pressure tightness	8 bar, no leak	Air pressure test on every lot
Cylindricity	0.2 mm	Dial indicator
Perpendicularity	0.1 mm	Dial indicator
Corner radius	R1.5 max	Radius gauge
Packaging	500 pcs per bag	Dent and crush protection, counted per bag

The nine-step process chain

Every step below carries a documented control: a first-off check, an in-process frequency, a reaction plan if it drifts, and a named owner. That structure — not the machine list — is what a PPAP reviewer actually reads.

#	Step	Equipment	Controlled characteristic	Frequency
1	Incoming tube inspection	—	Ø4 ± 0.05, Ø3 ± 0.05, HV 300 ± 50, mill certificate	AQL 0.15 dimensional / 0.40 visual, G-II
2	Tube cutting	Instrument lathe	Min 108 mm blank length	5 pcs first-off, 5 pcs every 2 h
3	Welding	Welding machine	Min 108 mm, R1.5 max, 8 bar tightness	5 pcs every 2 h, leak test each lot
4	Outside diameter grinding (subcontracted)	Centreless grinder	Ø4 +0 / −0.05	Supplier dimensional report per lot
5	Edge trim and chamfer	Instrument lathe	106 ± 0.5 mm, no burr at tube mouth	5 pcs every 2 h
6	Vibratory polishing	Vibratory bowl	4 h cycle, logged on the patrol record	Appearance AQL 0.40
7	Final quality inspection	—	100% visual; dimensional sampling	AQL 0.25 dimensional, every lot
8	Packaging	Poly bag, counting scale	500 pcs per bag, no dents or crushing	Count and label check per bag
9	Shipping documentation	—	Packing list, material ID, final inspection report, material certificate, SGS report	Every shipment

First-article verification: 45 checks, zero rejects

The submission run was checked on five pieces against all nine characteristics. Measured values, as recorded on the sample report:

Characteristic	Spec	Measured (5 pcs)	Gauge	Result
Length	106 ± 0.5	106.00 – 106.08	Caliper	Accepted
Outside diameter	Ø4 ± 0.05	3.97 – 3.98	Caliper	Accepted
Inside diameter	Ø3 ± 0.05	Go / no-go, all pass	Plug gauge	Accepted
Outside diameter (ground)	Ø4 +0 / −0.05	3.97 – 3.98	Caliper	Accepted
Hardness	HV 300 ± 50	318 – 339 HV	Hardness tester	Accepted
Pressure tightness	8 bar	OK on all 5	Air pressure rig	Accepted
Cylindricity	0.2	OK on all 5	Dial indicator	Accepted
Perpendicularity	0.1	OK on all 5	Dial indicator	Accepted
Corner radius	R1.5 max	OK on all 5	Radius gauge	Accepted

Note the direction of the diameter readings: 3.97 – 3.98 mm sits at the low end of a Ø4 +0 / −0.05 band. That is deliberate. The ground OD runs toward the low limit so the part still assembles after the welding heat cycle, which is where a tube like this normally grows.

The measurement system was validated, not assumed

A gauge R&R study was run on the length characteristic before the dimensional data was accepted: three appraisers, ten parts, three trials.

Component	Value
Equipment variation (EV, repeatability)	0.0055 mm
Appraiser variation (AV, reproducibility)	0.0007 mm
Gauge R&R (GRR)	0.0056 mm
Part variation (PV)	0.0287 mm
Total variation (TV)	0.0292 mm
%GRR of study variation	19.0%
%GRR of the 1.0 mm tolerance band	3.3%

19.0% of study variation lands in the 10–30% band, which is conditionally acceptable rather than ideal — and the honest reason is that part-to-part variation on this characteristic is small (0.029 mm) while a caliper repeats to roughly 0.006 mm. Against the actual tolerance, gauge error consumes only 3.3%, so the measurement is fit for the decision being made. Where a characteristic is tighter than the gauge, the answer is a different gauge, not a better spreadsheet: the Ø3 bore is checked with a go / no-go plug gauge for exactly that reason.

PFMEA: the highest RPN is 12

The process FMEA covers the same nine steps. Severity never exceeds 3, occurrence never exceeds 2, and detection never exceeds 2. Average RPN across the nine rows is 8; the highest is 12, on incoming material specification, cutting, welding, and subcontract grinding.

Those numbers are low for a structural reason rather than an optimistic scorer: every failure mode is either caught in-process at a two-hour interval or prevented upstream by material certification. The weakest link is not the shop floor — it is the subcontracted grinding step, which only sees a supplier dimensional report and an incoming AQL sample. That is a known exposure on this part family, and it is managed by supplier qualification rather than by inspection volume.

What the package actually contained — and what it did not

That distinction matters if you are comparing suppliers. “45 of 45 readings in tolerance” proves the process can make a conforming part on the day of the run. It does not prove the process holds that spread over a thousand pieces — only a capability study does. Any page that shows a Cpk without saying how many pieces and which characteristic it came from is giving you a number, not evidence.

What this means for your part

• Thin-wall stainless tube with a weld and a leak spec — the sequence that works is weld first, grind after, and hold the OD low. Grinding before welding means reworking a heat-distorted diameter.
• Tight bores need attribute gauges. A Ø3 ± 0.05 bore on a 106 mm tube is not a caliper job; go / no-go is faster and more repeatable than a number nobody trusts.
• Put the sampling level in the control plan, not in a meeting. AQL 0.15 dimensional and 0.40 appearance, at a stated G-II level, is a number your auditor can check.

Related reading

Other anonymised packages in the same format are collected under deep-drawn part case studies. If your part is a drawn shell rather than a welded tube, the process envelope and the tolerances that go with it are different — start there.

All figures on this page are taken from the production PPAP file for this part. Customer identity, part number, drawing revision, purchase order, supplier names and inspector names have been removed. Measured values are reproduced as recorded on the sample report.