Acceptance Class Selection for Electronics Manufacturing
Every board is inspected against something. If the criteria are not written down, the inspector applies personal judgement, and the result varies between shifts, sites and customers. An acceptance class is a way to state in advance how much deviation is tolerable, so that a joint is either accepted or rejected for a reason that both parties agreed on. This article explains what the classes change, what each step costs and how the criteria are built into the product documentation.
What an Acceptance Class Is Not
A class is not a quality grade for the product, and it is not a specification of reliability. It is a set of acceptance criteria for workmanship, describing how a joint, a pad or a coating must look and how much deviation is allowed. A class 3 assembly is not inherently more reliable than a class 2 assembly of the same design; it is inspected against tighter criteria and usually built with tighter process controls.
The distinction matters because it prevents a class being used as a marketing description. Choosing a higher class changes the cost of inspection and the amount of rework, and it only improves reliability when it is combined with a process that can actually deliver the tighter criteria.
What Changes Between Classes
The classes differ in the amount of deviation they permit. Higher classes require more complete solder fill in a plated hole, tighter limits on voids, smaller permitted misalignment, more complete fillets and stricter rules on the appearance of a wet joint. Cosmetic features that are acceptable in a general product class may be rejectable in a high reliability class.
The criteria also change how conditions that are not defects are treated. A mark, a discolouration or a slightly non uniform coating may be acceptable at one class and not at another, which matters because those are the conditions that generate the most disagreements between a supplier and a customer. The point of writing inspection criteria in advance is that a disagreement becomes a reference to a document rather than a difference of opinion, and the document can be updated when experience shows that a criterion was misjudged.
Cost of Moving Up a Class
Moving up a class changes three things: the scrap rate, the inspection effort and the process capability required. A design that is comfortable at a general class may need a longer reflow dwell, a higher paste volume or a different finish to meet the tighter fill and void limits, which changes the cost before any board is inspected.
The inspection cost rises as well, because more criteria require measurement rather than visual judgement. Void limits in particular require X-ray or cross sectioning rather than a look at the joint, and a class that specifies a void limit therefore implies an inspection method and a sample size. The total cost should be estimated before the class is chosen, not discovered afterwards.
Void and Fill Limits
Voids in a solder joint are the criterion that causes the most confusion. A void is a gas pocket inside the joint, and its effect depends on where it is and how much of the cross section it occupies. A void in a thermal pad has a different consequence from a void in a signal joint, and a void at the interface behaves differently from one in the bulk.
The limit should therefore be written with a method and a location, not only a percentage. Which joints the limit applies to, how the void area is measured, what sample size is used and what the disposition is when the limit is exceeded all belong in the specification. The reasoning that connects a limit to joint performance is the same as the logic in solder joint acceptance criteria.
Writing the Criteria Into the Drawing
The class belongs on the fabrication and assembly drawing, together with any exception that applies to the product. A board that carries a power device with a void limit different from the general class should state that limit at the feature, and a connector that is exempt from a cosmetic criterion should be identified. Without that, the inspector has to interpret the requirement.
The drawing should also state the inspection method where the criterion implies one. A requirement for a fill percentage implies a cross section, a requirement for a void area implies X-ray, and a requirement for a coating thickness implies a measurement. Stating the method removes the argument about whether the criterion was applied correctly. The data package conventions that keep this information together are described in PCB fabrication notes.
Training, Consistency and Audit
Criteria only work if they are applied consistently. Inspectors need training against the actual standard rather than a summary, and they need a reference set of samples that show the accept and reject boundary. A reference set is also the fastest way to align two sites, because it removes the differences in interpretation that written criteria alone cannot resolve.
Auditing closes the loop. Reviewing a sample of accepted and rejected boards against the standard shows whether the criteria are being applied as written, and it identifies the criteria that cause the most disputes so that they can be clarified. The same review applied to the process data, as described in first pass yield analysis, shows whether the process is capable of the class that was specified.
Choosing the Class in Practice
The choice should follow the consequence of a failure. A product where an intermittent joint causes a service call, a safety concern or a recall justifies a tighter class and the process controls that go with it. A product where a defect is detected at final test and corrected has more room.
The gopcb engineering team recommends deciding the class at the design stage and reviewing it with the customer, because the decision changes the layout, the finish, the stencil and the process. Where a product sits between two classes, the useful approach is to identify the specific criteria that matter and specify them individually, rather than adopting a whole higher class and paying for requirements that the product does not need.
Aligning Criteria Between Sites and Suppliers
A product built at two sites will not be inspected identically unless the criteria are defined the same way. The written standard is necessary but not sufficient, because the same wording can be interpreted differently when it is applied to a real joint. A reference set of samples that shows the boundary, together with a periodic cross check where both sites inspect the same boards, is what keeps the two results comparable.
Supplier alignment works the same way. Where a supplier is measured against criteria that were not agreed, the result is a stream of rejected lots that the supplier cannot correct. The criteria should be part of the purchase specification, and a periodic review of rejected samples with the supplier converts a rejection into a process correction. That review also identifies the criteria that are not practical, which is how a specification improves over time rather than becoming a source of permanent friction.
FAQ
Does a higher class guarantee a more reliable product? No. It specifies tighter acceptance criteria and usually tighter process control. Reliability depends on the design and on the process delivering those criteria.
Can different classes apply to different features on one board? Yes, and it is common. The class statement should identify the features that are exceptions and the criteria that apply to them.
What happens if a board fails a criterion? It is dispositioned as rework, use as is with documented approval, or scrap. The specification should state who is authorized to approve a use as is decision.
