Keep Engineering-Trial Assemblies Out of Unrestricted Production Stock

An experimental assembly can pass a normal test while still containing an unapproved material, process setting, or modification. Its trial identity must survive testing, storage, and handling. A green test label should not automatically turn engineering work into shippable inventory.

Define the trial population before it starts

Assign the experiment an identifier and list the board serials or other traceable units. Record the intended departure from production: alternate component, process window, rework method, firmware, or material condition. Include a disposition owner and define whether any route to production release is permitted.

TI’s LMX2615 evaluation-board description explicitly distinguishes its installed engineering sample from devices that have completed the full space production flow. It is a concrete example of why successful evaluation behavior and production qualification are separate attributes. Each trial’s actual restrictions must come from its own approved plan.

Carry status through every handoff

• Mark the physical units and their storage containers with the trial identity.
• Assign a corresponding status in the inventory or tracking system.
• Link process and test results to both the unit and the experiment.
• Prevent routine dispatch from treating a trial pass as unrestricted acceptance.
• Reconcile remaining boards and unused material when the trial closes.

Use test-recipe identification so a standard program run and an experimental program run remain distinguishable. A passing functional check may demonstrate one objective of the trial without addressing its material or workmanship departures.

Approve conversion explicitly, if it is allowed

For example, ten boards use a proposed alternate coating and all pass a short functional test. That result does not establish the environmental performance required for the production product. Keep them in trial status until the planned evidence is evaluated and an authorized disposition is recorded. Do not solve a shipment shortage by changing only their inventory label.

If the trial changes an assembly configuration, document the resulting as-built state through the work-in-progress change boundary. Any permitted conversion should identify required rework, retesting, marking changes, and the exact units covered.

Close the experiment with a count of units retained, consumed by testing, converted under approval, returned, or otherwise disposed of. Retain trial results even when the proposed change is rejected; they explain why those boards cannot be treated as ordinary stock and help prevent the same ambiguous material from resurfacing in a later build.