Assembly DFM Review Agenda

A manufacturability review is the last cheap opportunity to change a design. After the layout is released, a change costs a revision, a new stencil, a new program and possibly a scrapped panel; before release, it costs a conversation. The difference between a review that finds problems and one that produces a list of general observations is almost entirely the agenda: specific questions about specific features, asked in a defined order.

Why a Structured Agenda

An unstructured review drifts towards whatever the most vocal participant is worried about, and the topics that nobody raises are exactly the ones that cause problems later. A structured agenda walks the design from the panel inward: the frame and the rail, the placement of the components, the soldering process, the test and inspection access, and the documentation. Each section has a small number of questions, and each question has a person who can answer it.

The agenda also makes the review repeatable across products. The same questions asked of every design produce a comparable record, and the recurring findings identify the design rules that need to be updated. That feedback loop is where most of the value accumulates over time. Fabrication checklist practice follows the same principle.

Who Should Attend

The review needs the design engineer, the process engineer who will build it, the test engineer, the quality function and, where the product is complex, the person who will handle the material and the kitting. A customer representative is useful when the product is built to a customer specification, because some findings require a customer decision rather than an internal one.

Attendance should be by role rather than by seniority, and the meeting should be short. A review that runs for three hours will not be repeated on the next product, and the value comes from repetition. Where a topic needs deeper analysis, it should be taken away as an action rather than resolved in the room. Quality criteria should be available to the group so that the discussion is grounded in the standard.

Panel, Rail and Frame

The first section covers the mechanical interface. How many boards fit on the panel, and is the utilisation acceptable? Is the rail wide enough to be handled and does it match the standard frame? Are the tooling holes present, correctly sized and in the standard position? Are the fiducials present, on both the panel and the boards, and clear of other features?

The separation method belongs here too. Which method will be used, and does the layout support it? Are the tabs or the score line positioned away from critical components, and is the resulting edge acceptable for the product? A panel that is difficult to separate will produce stress damage on every unit, and the decision is much cheaper to make now than after the first build. Panel yield analysis quantifies the trade.

Placement and Process

The placement section covers the process the product will actually run on. Are all the packages within the capability of the machines, and are the nozzles available? Are there components that require a special feeder or a different tape format? Is the orientation of polarised parts unambiguous in the documentation, with a consistent convention across the board?

Then the constraints: are there components placed where they will interfere with a support pin, a conveyor edge or an existing fixture? Is there enough clearance for the placement head to reach a tall neighbour? Where a part is placed close to the rail or to a break line, is the risk acceptable? These questions are answered by looking at the panel against the machine, not by reading the layout alone.

Solder and Thermal

The soldering section covers the process window and the thermal design. Does the assembly have a combination of thermal masses that the profile can accommodate, and are there any components with a temperature limit that the process will exceed? Where a part is temperature sensitive, is the protection method defined and is the sequence set so that it is not exposed twice?

For through hole parts, are the thermal reliefs specified, and are the hole sizes appropriate for the fill requirement? For area array packages, is the pad geometry suitable and is the paste volume achievable with the stencil thickness? For a mixed technology board, are the wave and reflow steps sequenced in a way that protects the joints made first? Each of these is a specific question with a specific answer.

Test and Inspection

Test access is the section that most often produces findings. Are the test points present, accessible and of a size the fixture can probe? Does the design allow the fixture to support the board, and is there room for the probes between tall components? Has the test strategy been defined, and does the layout provide what it needs?

Inspection access matters as much. Can optical inspection see the joints that need checking, or does a shield or a connector obscure them? Is X-ray required, and is the equipment available? Where a joint cannot be inspected, has that been accepted deliberately? Documenting an accepted blind joint is much better than discovering it during the first build.

Documentation and Readiness

The final section confirms that the build package will be complete: the bill of materials with approved alternates, the assembly drawing with orientation marks, the stencil and tooling data, the profile, the inspection criteria and the packaging specification. Each item should have an owner and a date, and the list should be reviewed again at the first build.

Material readiness belongs here as well. Are all parts available, and are the long lead items ordered? Are there parts with a single source, and has an alternate been assessed? Where the answer is no, the risk should be recorded and accepted explicitly rather than discovered during the ramp. Design release practice provides the framework for the final check.

Action Tracking

Every finding should leave the meeting with an owner and a date, and the list should be reviewed in the next session. Findings that are not tracked are findings that were interesting rather than useful. Where a finding is rejected, the reason should be recorded, because the same question will be asked again on the next product and the answer should be available.

The tracker also feeds the design rules. A finding that recurs across products is not a design error but a gap in the rules or in the checklist, and closing that gap prevents the discussion from happening again. That is how a review process improves the design process rather than merely catching its mistakes.

FAQ

When should the review be held? Before the layout is released for fabrication, and again briefly before the first build to confirm that the actions were closed.

How long should it take? An hour for a moderately complex assembly. Longer meetings are usually a sign that the design is not ready for review.

What if the team disagrees? Record the disagreement and its consequences, take the data away, and decide. A review is not a vote.

How do we know the review was useful? By the number of findings that were closed before the first build, and by the reduction in findings on the next product.