A first-article report is the first hard evidence that a new supplier can make your part. A weak one looks complete: every row ticked “OK”, a signature at the bottom. A strong one lets an engineer reconstruct exactly what was made, how, and how it was measured.
First, are these really first articles?
First articles should be made with the production die or tooling, on the production press, from production-grade material bought from the intended source, by the people who will run the job. Prototype or soft-tool parts are useful for development, but they don’t count. If any of those four conditions is different, the report describes a different process.
The header
Every report should identify:
- part number and drawing revision,
- supplier and production site,
- die and press identifiers,
- material heat or coil number,
- date made, who measured, and the equipment used.
Without the die, press and coil, a later problem can’t be traced back to what changed.
How many parts, measured how far
Two different things are measured. A full layout checks every characteristic on the drawing, including title-block dimensions. A larger sample checks only the critical (CC) and significant (SC) characteristics from the requirements matrix. DSQP scales both with the part’s risk level:
| Risk level | Full layout | Parts measured on each CC and SC | Material evidence | Assembly trial at buyer |
|---|---|---|---|---|
| C | At least 1 | 3 | Mill test report | Recommended |
| B | At least 2, across setups or streams | 5 | MTR; thickness measured | Required |
| A | At least 3, covering each cavity, stream or setup | 10 | Traceable evidence; lab test if required | Required, with the buyer’s operator |
The results table
For each characteristic and each sampled part, the report should show the requirement and limits, the part identifier and the actual measured value, the method or gauge, and whether it conforms. “OK” is not a value. Neither is a column of numbers that all equal nominal.
Material
Check the mill test report against the specification on the drawing, not just its presence. An MTR describes the coil as shipped from the mill; it is not proof of every finished-part property. Where the specification, the customer or the risk review requires it, an independent laboratory test is ordered.
When something fails
A failure is information, not automatically a disqualification. Confirm the measurement first, then investigate and correct with a documented corrective action: contain, describe, find the cause with evidence, correct, verify by re-measuring with the same method, and update the die record and setup sheet. Corrections that change design, material or functional geometry need the buyer’s engineering approval.
Red flags in a first-article package
- ticks or “OK” instead of measured values,
- no die, press or heat/coil identifiers,
- parts made on prototype or soft tooling,
- measured with a different gauge from the one production will use,
- drawing revision missing or out of date,
- failures with no corrective-action record, or a correction with no re-measurement.
Atlas Insights are general technical guidance. Customer, safety and regulatory requirements, and your engineer’s decisions, always govern.