MANUFACTURING / JAPAN SOURCING
PCB prototypes to production: define what each test result proves
A working prototype shows what happened on that board under those conditions. Turn the intended production checks into agreed methods, limits and records before asking a factory to repeat them.
A demonstration leaves questions for the production line
Imagine a controller-board prototype that powers up and sends one status message on a developer’s bench. Purchasing now asks an assembler for a production quotation with “test as per sample.” This fictional case leaves the assembler without an agreed response to a failed output, a different firmware image or a missing test connection. The working board is useful evidence, but it is not a complete production test specification.
Begin with the functions and assembly faults the production test must address. Keep design validation, production inspection and the finished machine’s acceptance distinct. A development trial may explore a wide operating range; the line needs a defined, repeatable decision for the units it produces. Decide which evidence belongs at each stage rather than assuming that one successful demonstration covers them all.
Choose a method for each question
Keysight describes in-circuit testing (ICT) for assembly faults and separate functional-test capabilities.[1] The methods answer different questions. For this board, ask how the proposed inspection will detect the faults and functions that matter, what electrical access it needs, and which items remain outside its scope. Owning a named tester does not establish coverage of your design.
NI identifies power, component operation and communication among functional-test needs, alongside cycle time and fit with manufacturing.[2] Apply those questions to the actual board. “Power on” may only mean that a light came on; if the required output and communication response were never stimulated or measured, that result says nothing about them.
For every required check, specify the stimulus or operating state, the measured response, the decision criterion, and the record to retain. Have the responsible technical team set the values and tolerances. Leave unresolved limits visible in the quotation; do not invent them from a prototype’s single reading.
- 1Board: revision, fitted parts and firmware identify the unit tested
- 2Fixture and station: access points and connected equipment define the setup
- 3Record: conditions, measured values and the decision travel with the unit
Illustrative controller board and test station. Lines show test access and the link to a record, not an electrical connection plan. The blank record is not a passed result.
Make the prototype-to-line handover concrete
The following example is a planning sheet for the fictional controller, not a completed test procedure. It distinguishes the bench observation from the evidence still needed for production. The final scope may include visual or other inspection as well as electrical testing; select those methods for the board and its risks.
| Question | Prototype evidence | Production handover still needed |
|---|---|---|
| Power behaviour | Prototype evidenceThe sample started on the bench | Production handover still neededDefined supply and load conditions; readings, limits and sequence |
| Outputs and communication | Prototype evidenceOne status message was observed | Production handover still neededRequired functions, stimuli, measured responses and exclusions |
| Board and software identity | Prototype evidenceThe developer knows the sample | Production handover still neededBoard and BOM revision, firmware identity and programming record |
| Physical test access | Prototype evidenceTemporary bench leads were used | Production handover still neededApproved access points, fixture contacts and support, station configuration |
| Failed results | Prototype evidenceThe demonstration worked once | Production handover still neededFailure record, segregation, authorized repair and retest rules |
| Repeatability and release | Prototype evidenceNo line trial has been agreed | Production handover still neededPilot units, comparison with agreed reference evidence and release owner |
The fixture is part of the measurement conditions
Ask how the board will be supported, contacted and connected without altering the condition being assessed. Identify the test points or connector pins from the released design data, the required power and loads, and the instruments and software used. These are engineering inputs to a fixture, not choices to leave to a photograph of the prototype.
Keep the fixture drawing or configuration, test-program revision, firmware and approved limits with the setup. Agree how contact wear, maintenance and a fixture or software change will be checked before affected results are accepted. A reference unit can help compare station behaviour, but it must have an identified, supported status; a unit that merely worked once does not establish every pass/fail limit.
Use the general inspection-fixture article for support and access questions. This article adds the board’s electrical stimulus, software and result traceability; mechanical repeatability alone cannot demonstrate circuit function.
Quote the work needed to produce and keep the evidence
Ask the assembler to separate test development, fixture and equipment, programming, per-unit testing, records, and any later maintenance or transfer work. Specify ownership and handover of the files, equipment and results as commercial terms rather than assuming they are included in an assembly price.
Before the planned production release, run the agreed trial and reconcile the results with the approved method. Record deviations and the authorized scope, including any checks deferred to another stage. Keep individual unit or lot identification, the applicable data revisions, results and retest history to the level agreed for this product. A successful retest should not silently erase the initial failure.
Bring HAM the board revision, intended functions, quantity, destination and the test information already available. We can help put the open questions to Japanese suppliers so that the quotation describes the work and evidence still needed. The product’s technical team approves the method, limits and release.
Questions to send your supplier
Open supplier questions
Board and BOM revision; fitted variants: Firmware image/version and programming requirements: Intended functions, use and destination: Prototype records: sample identity, setup, conditions and results: Production checks: stimulus/state, response, method and exclusions: Limits and tolerances: approved values or unresolved items: Test access, fixture support, power, loads and instruments: Station configuration, program revision and retained records: Planned quantities, cycle-time needs and pilot-run scope: Unit/lot identification, failed results, repair and retest rules: Development, tooling, test and maintenance scope to quote separately: Ownership, transfer files, release owner and remaining decisions: Please propose a test scope and identify every assumption, missing input and check not included.
Sources
- Keysight — In-circuit and functional test for manufacturing
Read HTML descriptions of ICT, boundary scan and functional test. System capabilities are not coverage guarantees for a particular board; no vendor percentage or test limit adopted.
- NI — Electrical and DC functional test station
Read the requirements section on power, component operation, communication, cycle time and manufacturing fit. No product selection, performance guarantee or brochure/PDF review.