Engineering Change Impact Assessment

An engineering change is never only the change that was requested. A component replacement alters the bill of materials, the placement program, the kit, the test limits and possibly the stock that is already on the shelf. A revision to a drawing alters the operator instruction, the inspection criteria and the traceability record. Assessing those effects before the change is released is what prevents a revision from creating a mismatch that surfaces weeks later.

What Triggers an Assessment

Any change that alters a released document should trigger one: a design correction, a component substitution, a process improvement, a supplier change, a customer request or a regulatory update. The trigger should be formal, because changes that begin as a conversation on the shop floor are the ones that escape the process entirely and appear later as an unexplained difference between two builds.

The assessment also applies to changes that appear trivial. A resistor value that improves a tolerance, a connector that is now available in a different colour or a firmware version that fixes a minor bug all have downstream effects on documentation, stock and test. Treating them as trivial is how the mismatch begins. Document control provides the framework that makes the trigger automatic.

Drawings and Documentation

The first pass identifies every document affected by the change: the schematic, the layout, the assembly drawing, the bill of materials, the work instruction, the inspection criteria, the test program, the packaging specification and the customer documentation where the change is visible to them. Each should be marked as requiring an update, with an owner and a date.

The critical question is whether the documents will change together. A drawing released a week before the corresponding work instruction creates a window in which the line builds to two different sets of instructions, and the product built in that window is of uncertain conformance. Where a change cannot be released simultaneously, the sequence should be defined and the interim state documented rather than left to chance. Sample retention is useful here, because it preserves the evidence of what was built before the change.

Material and Stock

Material impact has two halves: what will be purchased in future and what is already on the shelf. Obsolete stock has a cost, and the decision to scrap, rework or use it up must be made deliberately, with a limit on the quantity and a record of which builds received it. Where the old material is consumed without a record, a later defect investigation has no way to distinguish the affected units.

The availability of the new material is a schedule risk rather than a technical one, but it belongs in the same assessment. A change that is technically sound and cannot be supplied for four months should be released in a way that matches the supply. Where the material is on allocation, the assessment should say so, and the plan should account for the possibility that only part of the quantity arrives. Incoming inspection should be told what to expect, particularly where the new part looks similar to the old one.

Tooling, Fixtures and Programs

Tooling changes are often overlooked. A new package may need a different nozzle, a different stencil aperture or a different support plate; a moved connector may invalidate the test fixture; a component with a different height may interfere with a shield or a coating mask. Each of these has a lead time, and each of them can stop a line that is otherwise ready to build the new revision.

Programming changes belong in the same list. Placement programs, printer recipes, oven profiles, coating paths and test programs all reference the product, and a change that is not reflected in the program produces a defect that no operator can catch. The assessment should state which programs are affected and who is responsible for the update and its verification.

Test and Inspection Impact

A change to a component or a circuit can invalidate a test limit. A new part with a wider tolerance may fail a limit that was set for the old one, and the failure will look like a defect rather than a measurement issue. Where the change affects a functional parameter, the test limits should be reviewed with the same care as the design, and the reason for any adjustment recorded.

Coverage should also be reviewed. A change that moves a component under a shield or behind a connector may remove an optical inspection capability, and a change that adds a test point may enable a simpler check. Recording the effect on coverage keeps the inspection plan aligned with the product. Inspection standards should be updated at the same time as the drawing.

Approval and Communication

The approval should be given by somebody who has seen the whole assessment rather than only the technical change. A short review with the affected functions represented, and a decision recorded with the conditions attached, is enough for most changes. Where the change affects a customer requirement or a regulated characteristic, the customer notification route should be part of the same decision.

Communication is the step that most often fails. The line needs to know that a change is coming, when it is effective and how to tell the old units from the new. A brief note at the workstation, an updated work instruction and a traveller that records the split point are simple measures that prevent the build from being a mixture of two revisions with no way to tell them apart.

Effectivity and Split Points

Effectivity defines when a change takes hold: at a date, at a work order, at a serial number or at the exhaustion of existing stock. The choice matters because it determines how the traceability record must be kept. A change effective at a serial number requires the serial to be applied before the change, while a change effective at the exhaustion of stock requires the remaining quantity to be known and recorded.

Where a build contains both revisions, the split point should be recorded on the traveller and in the batch record, with the two groups identifiable. That allows a defect found later to be attributed correctly, and it allows the older revision to be analysed separately if the change was made to solve a problem. Assembly traceability is what turns that record into an answer.

Verification After the Change

The assessment should end with a plan to verify the change rather than with the release of the documents. A first article on the new revision, a check of the affected inspection criteria and a short review of the yield over the following builds confirm that the change did what it was intended to do. Where the change was made to fix a defect, the verification is the evidence that the fix worked.

Where the change was made for a commercial reason, the verification is the confirmation that nothing else moved: the yield, the cycle time, the test time and the defect pattern should be compared with the baseline. Without that comparison, a change that introduced a new problem can go unnoticed because the original issue was resolved. Batch records provide the data for the comparison.

FAQ

Who should own the assessment? The engineer responsible for the affected area, with contributions from production, quality and purchasing. One owner should hold the record.

How long should an assessment take? Hours for a simple component change, days for a design change. If it takes longer, the process is too heavy and will be bypassed.

What about changes that improve yield? They need the same assessment. A process improvement that alters a parameter belongs in the same record as a design change.

How is the change verified? By a first article on the new revision and a comparison of the key metrics with the baseline over the following builds.