Intercontinental Casting — Integrated Supply Chain Platform for Non-standard Casting Parts

Why Mass Production Differs From Your Approved Sample

Publication Date : 2026-11-13 03:10:31

Why Mass Production Differs From Your Approved Sample

Your approved sample and your production shipment were made by the same supplier, on the same drawing — yet they can differ in wall thickness, hole position, surface finish and alloy. This article explains the four mechanisms behind that drift (tooling wear, process parameter drift, fixture and operator variation, and inspection dilution), gives the numeric tolerances and checkpoints worth writing into a control plan, and shows how to hold a supplier to them.

The Sample Was a Snapshot; Production Is a Trend

An approved sample is usually made under favorable conditions: a fresh or newly repaired tool, a short run, a seasoned operator, and extra inspection attention because everyone knows the part is being judged. Production is the same process run for thousands of cycles, on three shifts, with tools that are no longer new. The sample proves the design is feasible. It does not prove the process is stable, and those are different claims.

This is why a first article that matches the golden sample is not the end of the story. The useful question is not "does part one match?" but "what keeps part 5,000 inside the same window?" Answering that requires knowing which variables move over a production run and how much they are allowed to move before someone adjusts the process.

Typical dimensional drift between sample and mid-run production, by feature type (mm)Hole position0.12Wall thickness0.25Flatness0.18Draft-angle-dependent OD0.09

Tooling Wear: The Slow, Predictable Drift

Tooling wear is the most predictable source of drift, which makes it the easiest to plan for. In die casting, a die that starts with a 1.5 mm wall can thin toward 1.3 mm as the cavity erodes and the die heats and stresses over tens of thousands of shots; die spray and thermal fatigue also change surface texture. In sand and investment casting, pattern wear and wax-pattern shrinkage variation shift dimensions gradually. In machining, carbide inserts wear measurably within a few hundred parts — a boring bar that holds ±0.02 mm when fresh may drift toward ±0.05 mm before the next index.

The practical response is a wear allowance and a documented tool-life limit. Build the tool to the high side of the tolerance band on features that will wear toward the low side, define at what shot count or part count inserts are changed, and record actual measurements at the start, middle and end of each tool life. If the supplier cannot tell you the expected tool life in shots or parts, that is a gap in their process knowledge, not a detail.

  • Wear allowance: build wearing features 30–60% into the tolerance band, not at nominal
  • Tool life: record expected shots or parts per tool, and the trigger for repair
  • First-off, mid-life and end-of-life dimensional reports on the same drawing
  • Insert change interval stated in parts, with a first-piece check after every change
  • Die or pattern repair log with before/after measurements on the critical features

Process Parameter Drift: Heat, Pressure, Time

Even with a perfect tool, the process window moves. Metal temperature in a die-casting shot sleeve, sand moisture and compactability in green sand, wax injection pressure, melt chemistry, quench delay after solution treatment, and oven temperature uniformity all shift through a shift and across seasons. A 10–15 °C change in die temperature can shift shrinkage enough to move a critical dimension by a few hundredths of a millimeter; a humid summer week can change green sand strength and produce scabbing or dimensional scatter that never appeared during sampling.

Control plans should list each of these parameters with a nominal, a tolerance and a check frequency — not just the final part dimension. Checking the part tells you that you have drifted; checking the parameter tells you why, before the parts are scrap. Ask for the parameter log from a normal production week and compare it against the settings used for your sample. If the sample was run at 680 °C and production runs at 700 °C to improve fill, the sample is no longer representative.

Where dimensional nonconformities typically originate in a casting process (share of causes)30%28%18%14%10%Tooling wear (30)Process parameter drift (28)Fixture and setup variation (18)Operator method (14)Incoming material variation (10)

Fixtures, Operators and Inspection Dilution

The sample may have been machined in a single fixture by one experienced operator. Production spreads across fixtures, shifts and people. Fixture wear, clamp force and locating-pin condition introduce position error; different operators load parts differently, deburr differently and judge surface finish differently. A practical control is to require the same fixture ID and the same torque or clamp settings to be recorded with each lot, and to requalify fixtures on a fixed interval.

Inspection is the other half of the problem. Sampling plans that look rigorous on paper often mean one part per shift measured by whoever is free, with a gauge that has never been checked against a master. Define the gauge, the frequency, the sample size and the reaction plan in the control plan, and require gauge repeatability and reproducibility checks before the run starts. When a dimension is out, the plan should say who stops the line, what is quarantined, and how far back the suspect parts go — 100% inspection of the last lot is not a reaction plan.

Writing Drift Control Into the Purchase Order

You do not need a new quality system to close this gap; you need a control plan attached to the PO and a first-article report that covers more than part one. Specify the critical characteristics with their tolerances, the process parameters that affect them, the check frequency, the gauge to be used, and the tool-life and fixture-requalification intervals. Require a capability study (Cpk) on the critical dimensions from a run of at least 30 consecutive parts, and require the supplier to notify you before changing a tool, a fixture, a sub-supplier or a process parameter outside the agreed window.

Then verify with data, not promises. Ask for the dimensional report at first-off, mid-run and end-of-run of the first production order, and spot-check incoming parts against the golden sample on the characteristics that matter. A supplier who can show you the trend is a supplier who is watching it. One who can only show you the sample is asking you to trust a snapshot.

Frequently Asked Questions

If the first article passes, why do later parts fail?

A first article is usually made with fresh tooling, settled parameters and extra attention. Wear, thermal drift, fixture condition and operator variation accumulate over a run. That is why the control plan needs mid-run and end-of-run checks, not just a first-off report.

How much dimensional drift is normal before I should be concerned?

It depends on the feature and process, but as a working rule, drift of more than about 30–50% of the available tolerance band on a critical dimension means the process is not in control and should be investigated before the next lot ships.

What should I require in the control plan for a casting order?

The critical characteristics with tolerances, the process parameters that affect them with check frequencies, the gauge and its calibration status, tool-life and fixture-requalification intervals, a Cpk study on at least 30 consecutive parts, and a written reaction plan for out-of-tolerance results.

Looking for a reliable casting supplier? Request a quote from Intercontinental Casting — our engineering team replies within one business day.