IPC-A-610D for AOI: Integrated Verification Enables IPC-compliant PCB Inspection

Estimated reading time: 7 minutes

IPC-A-610 is the universally accepted standard for electronics assembly defects. In this article Peter Krippner, Viscom AG, describes how IPC-A-610D is incorporated into post-reflow AOI systems to improve defect detection, reduce false calls, and eliminate missed defects during SMT assembly.

During manufacture of electronic assemblies, quality standards are an ongoing concern. Systems for automatic optical inspection (AOI) are proven to secure product quality. The guidelines in IPC-A-610D are frequently drawn upon to evaluate and classify potential production defects. When these guidelines are used on an AOI system, several aspects need to be considered to ensure IPC-compliant inspection.

Production defects during the automated assembly and soldering of electronic assemblies are unavoidable. AOI systems are deployed increasingly to ensure that these defects are detected and to further optimize the SMT assembly process. Frequently, AOI systems are situated after the solder oven for post-reflow inspection, because this inspection stage presents the greatest potential for improving quality.

AOI system performance can be evaluated by defect detection, false alarms, and missed defects. Recurring themes such as inspection depth, or inspection features and their definition, come to the fore when acceptance criteria are specified. When specifications are described, exact defect definitions -- which defects are to be inspected for and what exactly a defect is -- are crucial in preventing later misunderstandings.

Here, IPC-A-610D (IPC), a "collection of visual quality acceptability requirements for electronic assemblies" (IPC, 1.1 Overview) has established itself as foundation and reference. IPC is the sole reference when inspection depth and features are specified, e.g.: "Inspection scope, inspection depth and inspection features are to be covered according to IPC-A-610D (Class 3)."

IPC-A-610’s pertinence for post-reflow inspection also suggests that it refers exclusively to the condition of soldered assemblies. In this article, I discuss IPC-A-610D applicability to automatic optical post-reflow inspection.

Overview: IPC-A-610D

IPC-A-610D’s collection of visual quality acceptability requirements for electronic assemblies subdivides assemblies into three classes, each with its own specific criteria (IPC, 1.4.1 Classification):

• Class 1 – General Electronic Products
• Class 2 – Dedicated Service Electronic Products
• Class 3 – High Performance Electronic Products

Among the experts, the introduction of a Class 4 is under discussion; still, the just released IPC-A-610E continues to list three classes.

The criteria for each class are then presented in four stages (IPC, 1.4.2 Acceptance Criteria): 

• Target Condition (ideal, to aim for)
• Acceptable Condition (permissible)
• Defect Condition (impermissible)
• Process Indicator Condition

The “ideal state” cannot always be achieved, but “acceptable condition” is sufficient. Process indicators comprise a special condition. They are not actual defects, yet do not fully satisfy the acceptance criteria. If an increase in process indicators is observed, the process should be analyzed to improve production.

The general criteria for solder connections are dealt with in Chapter 5 of the IPC manual. Chapter 8 covers surface-mount assemblies, where 8.2 is an extensive treatment of SMT solder connections.

 Figure 1. IPC-A-610D, 8-20: End joint width/displacement, chip component.

The acceptance criteria are very clearly illustrated, as the following example of a chip component (Figure 1) indicates. Due to space constraints, the criteria from 8.2.2.3 are only presented as excerpts:

Figure 1 describes the permissible boundaries of end width for Class 3: "The end width of the solder joint (C) is at least 75% of the component connection width (W) or the connection surface width (P), whichever is less."

It should be noted that the IPC-A-610 D describes criteria for visual inspection and presents standards for "visual inspection" (IPC, 1.2 Purpose). "Automatic inspection technology (Automatic Optical Inspection (AOI) represents a practical alternative to visual inspection..." (IPC, 1.6 Inspection Methodology). This means that the IPC standard was explicitly conceived for human visual inspection.

IPC 610 A’s Transferability to AOI

To what extent can IPC acceptance criteria for electronic assemblies be directly transferred to AOI? The ideal prospect in many cases is to adopt the dimensions for permissible acceptance conditions displayed in the illustrations as threshold values for AOI. Insofar as an AOI records proportions on the assembly geometrically, comparison with the adopted thresholds delivers the inspection result.

 Figure 2. IPC-A-610D, 8-50: Solder joint length on a Melf.However, one aspect deserves consideration here. As explained above, IPC is conceived as a tool for human visual inspection, with AOI as a practical alternative. Therefore, the threshold values are approximately quantified, as in the 25% increments for evaluation of chip components. This permits a subjective comparison of the schematic depiction with the actual assembly and the qualitative evaluation of visible features; finer gradations could hardly be implemented.

As well, not all dimensions are provided with thresholds, such as in the definition of solder joint length D on cylindrical end cap connections (Melfs). "Visually good wetting" (8.2.3.4) is listed. Certain components, like QFN/MLFs (Table 8-13), receive an even more rudimentary description.

A further example is the handling of "combined conditions." This term refers to the simultaneous occurrence of deviations for at least two features. IPC acceptance criteria are defined individually for each single feature; the large number of possible feature combinations prevents a complete definition (IPC, 1.4.2.5).

 Figure 3. IPC-A-610D, 8-51, Soldering on a Melf.

When defect features occur in combination -- e.g. insufficient solder on a displaced component -- a defect state can be present even though the individual features are within tolerances. This compounded defect can be detected by human visual inspection, implying that the human factor is critical in implementing IPC-A-610.

Simple transfer of available IPC-A-610 threshold values to AOI is insufficient for electronic assembly inspection. The IPC standard is a tool, yet does not replace the electronic assembly manufacturer's responsibility (nor that of the AOI manufacturer) for the quality of the finished product.

Concept for Applying IPC-A-610D to AOI

When IPC-A-610D is optimally incorporated together with AOI, it is used according to its original intentions.

 Figure 4. Concept for applying IPC criteria to AOI.

The basic idea is to draw upon IPC for final evaluation of the inspection results at the classification station (review of the inspection results provided by AOI), building a random sampling to verify good and bad examples for the AOI library. The following illustration (visually demonstrated in Figure 4) explains this correlation.

The AOI system delivers inspection results (image data) for review at the classification station. Generally, a zero-escape policy is already in place, meaning no defect may be overlooked (1). In actual practice, escapes sometimes occur, such as when defect conditions that were not previously considered crop up. A good AOI system must provide technology and methodology for such cases, to verifiably preclude repetition of a particular escaped defect.

The IPC acceptance criteria are taken into account by IPC-qualified personnel during inspection results classification and their ultimate evaluation (2). To guarantee the requisite level of knowledge, employees should receive regular training based on pre-classified images. This maintains and documents the quality of evaluations.

Using this methodology, AOI software evaluates classified AOI results to automatically sort the defect images. This provides a valuable verification data base containing good (pass) and bad (fail) templates (3). Verification databases can be supplied with inspection systems, and supplemented by the customer.

Feedback into AOI is provided by reliance on this IPC-compliant verification database to assess the AOI inspection library (4, 5). Corresponding off-line software automatically verifies whether the current parameter settings and thresholds correctly evaluate all template examples in the database (separation of classes).

This verification also produces explicit documents, so IPC-compliant inspection and zero defect escape can be proven during audits and other process control procedures.

Conclusion

IPC-A-610D has established and proven itself as a valuable tool in the manual visual inspection of electronic assemblies. A flexible concept for the application of IPC to AOI is required to fully exploit the performance capabilities of modern AOI systems.

 Figure 5. Separation of classes during integrated verification.

The assumption that available IPC thresholds need only be uploaded to the AOI and numerically compared with the geometric actual state is inadequate to support a zero-escape approach.

Integrated verification provides an image-based, classified verification database for manual review of inspection results, which can consider all influences and criteria for the good/bad decision. In addition to IPC, individual manufacturer requirements can also be incorporated.

Good software tools working automatically and off-line are critical to benefiting from this exceptionally valuable database in connection with an IPC-compliant inspection. Without additional time expenditure, AOI inspection library (and ultimately, the AOI itself) conformance to IPC standards is attained, securing zero-escape SMT inspection. Of course, since the IPC standard also describes many criteria that are not optically visible, this concept can also be fully applied to X-ray assembly inspection (AXI).

Peter Krippner, VP of assembly inspection, Viscom AG. Contact the company at http://www.viscom.com.Subscribe

Join the PennWell SMT Group on LinkedIn

Become a Fan on SMT's Facebook Page

Post your electronics manufacturing, SMT-related material to the #SMT community on Twitter. Use the #SMT hashtag.