Footprint Library: Lessons From the Last Product: A Design Review Input

The cheapest improvement available to most hardware programs is a short review of the previous product, and it is the one most often skipped because the schedule has already started. The findings are consistent, and they fall into a small number of categories.

Repeated Footprint Errors

The same footprint error appears in product after product, because the library is shared and the error was never corrected. A footprint that is wrong by a fraction of a millimetre passes on a coarse pitch package and fails on a fine one.

The remedy is a periodic audit of the library against the manufacturer’s drawings rather than against usage. An audit is tedious and it is done once, while the alternative is a defect that recurs indefinitely. Our component selection notes describe the package parameters.

Assumptions That Were Never Recorded

Every design rests on assumptions about current, temperature, tolerance and environment. Where they are not written down, the next design makes different ones, and the difference appears as a field failure rather than as a design error.

The record can be a single page. What matters is that a later reader can see what the design assumed and can check whether the assumption still holds.

Test Access That Disappeared

A design that was easy to test becomes difficult when a version is shrunk or a component is added. The nodes that lose their access are usually the ones that were added last to a crowded area.

The loss is silent, because the test programme is written after the layout. Reviewing the coverage as part of the layout review, rather than after the fixture is designed, is what prevents it. Our design for testability notes describe the provisions.

Grounding That Was Changed for Convenience

A ground plane that was continuous in the first version acquires a split in the second, added to solve a specific problem and left in place after the problem was solved differently. The split then interrupts the return path of a net that crosses it.

The check is to overlay the routing on the plane and to look for crossings. It takes minutes and it finds the problems that measurements find later.

Components That Became Obsolete

A product that has been in production for years accumulates obsolete parts, and each one is a redesign waiting to happen. The list is available from the purchasing record and it is rarely reviewed as a design issue.

Reviewing the list as part of the next revision allows the changes to be made together rather than one at a time under pressure. Our bill of materials notes describe the fields that track the risk.

Documentation That Diverged

After several revisions, the drawing, the bill of materials and the fabrication data may no longer describe the same product. The divergence is usually small and it is exactly the kind of small difference that causes a build to be wrong.

The check is to compare the three against each other rather than against memory, and to do it as one exercise rather than three.

Lessons That Were Not Written Down

A defect that is fixed in production is often not recorded in the design rules, so the next product repeats it. The record has to be made at the point of the fix, because the memory of the cause fades within weeks.

The practical mechanism is a design rule list that grows with each project. Each entry represents a defect that has already been paid for once. Our design review notes describe where the list is applied.

Reviewing the Previous Product

The review is short: the field returns, the production yield, the test escape rate, the obsolete parts, the engineering changes made during production, and the assumptions the design rests on.

Each of those produces a finding that can be applied to the new design at no cost, because the new design has not been drawn yet. Doing the same review after the layout is complete converts each finding into a change.

Process Control and Verification

Where the requirement is not written down, the shop supplies its own default, and the default is chosen for the process rather than for the design. The measurements that matter are the repeatable ones: conductor width and spacing, annular ring, finished hole size, plating thickness and surface finish are all verifiable on a coupon that travels with the panel.

Running a first article through the same checks as the production panel confirms that the two agree, and that comparison is the cheapest form of process control available at prototype stage. Keeping a sample from the panel turns a dispute into a measurement, because the same coupon can be re-examined by both parties without rebuilding the batch.

Checks Before Release

The checks that matter are the ones performed on the product rather than on a sample kept for the purpose, because a coupon that travels with the panel is the only evidence about that panel. Where a requirement can be measured, it should be measured at the point of manufacture and recorded against the board or the lot it applies to.

Verification and Records

A parameter that is set once and never re verified drifts, and the drift is usually discovered by a defect rather than by the record.

Points to Confirm at First Article

The tooling, the material and the profile form one system, and a change to any of them should be assessed against the other two before it is released. Where the process window is narrow, the measurement resolution has to be better than the window, or the data cannot distinguish a good part from a marginal one.

A record that identifies the operator, the date and the settings is worth more than a record that identifies only the result. The acceptance criteria should be written before the work starts, so that the decision is made by the specification rather than by the person inspecting.

Where an operation cannot be verified afterwards, it has to be controlled during the operation, and that control has to be visible in the record.

FAQ

How long should the review take? Half a day for a product that has been in production, most of which is spent collecting the data rather than analysing it. The findings usually number fewer than a dozen.

What if the previous product has no data? That is itself the finding. A product with no yield history, no return data and no engineering change record cannot be reviewed, and the new design inherits the same blindness.

What does gopcb provide for such a review? We provide the process records, the yield for each batch, the deviation reports and the coupon data for the previous product. Where the two products share a stack or a process we can compare the results directly.