A line can pass solder paste inspection and automated optical inspection, yet still contain a joint that cannot be trusted in service. This usually becomes apparent when assemblies use bottom-terminated components, BGAs, stacked packages, connectors with hidden solder areas, or metal shields that block the camera’s view. The visible side of the board may look clean while the actual solder interface contains voiding, bridging, insufficient collapse, or open connections.

X-ray inspection is needed alongside AOI and SPI when the quality decision depends on features that are hidden from an optical view. SPI confirms whether solder paste was deposited within an acceptable process window. AOI checks visible paste, component placement, polarity, markings, leads, and exposed solder fillets. X-ray should be added where those checks cannot verify the internal joint geometry or where the consequence of a hidden defect escaping is too high. The three methods are complementary rather than interchangeable.

Start with the joint, not with the machine list

The most reliable way to decide whether to introduce X-ray is to map each critical component to the evidence required for acceptance. A gull-wing lead with an exposed fillet can often be judged effectively by AOI, provided the optical system has adequate coverage and the inspection program is stable. A BGA solder ball underneath the package cannot. AOI may identify component shift, missing components, or obvious package damage, but it cannot directly see the condition of each hidden ball.

This distinction matters during process reviews. Teams sometimes add X-ray after a failure investigation because AOI was assumed to be a complete post-reflow inspection step. It is not. AOI is excellent at detecting conditions visible from above or from angled optical views. It does not replace an imaging method designed to inspect through the package body and board layers.

Component situations that normally justify X-ray

Not every PCB needs the same X-ray coverage. The decision becomes clearer when component architecture is considered.

BGAs, CSPs, and other area-array packages

These packages are the most familiar reason to use X-ray. Their solder balls sit between the package and PCB, leaving no exposed fillet for optical confirmation. X-ray can reveal ball-to-pad alignment, solder bridges between neighboring balls, missing or collapsed balls, and abnormal variations in joint shape. It can also identify whether a defect pattern is isolated or repeated across a package, which is useful when determining whether the cause may be stencil-related, placement-related, or thermal-profile-related.

For fine-pitch area arrays, inspection capability must match the feature size. Magnification alone is not enough. Image contrast, detector performance, geometric distortion, fixture stability, and the ability to separate adjacent balls all affect whether the image supports a dependable disposition decision.

QFNs, DFNs, LGA packages, and bottom-terminated devices

A QFN may have visible peripheral edges, but the main thermal pad and much of the electrical connection are underneath the body. AOI can often confirm that the package is present and oriented correctly. It cannot determine whether the central pad has spread correctly, whether solder has bridged underneath, or whether excessive voiding is present in a thermally important area.

This is especially relevant for power devices and parts that transfer heat through an exposed pad. A joint can remain electrically connected while having a solder distribution that raises thermal resistance or creates uneven mechanical support. X-ray does not automatically determine long-term reliability, but it provides evidence that optical inspection cannot supply.

Connectors, modules, shields, and stacked assemblies

Hidden solder joints are not limited to semiconductor packages. Some connectors have concealed terminations, and RF modules may include shielding or structures that restrict the line of sight. In multilayer or densely populated areas, the question is whether a defect can be observed clearly enough to make a decision. Where overlapping metal features make interpretation difficult, the inspection plan may need carefully selected viewing angles, region-specific recipes, or a review method capable of separating layers more effectively.

When SPI and AOI results point to a hidden-joint risk

X-ray is also valuable when upstream inspection shows a process condition that may not produce a visible downstream defect. A marginal SPI result does not prove that a joint will fail. Reflow can compensate for some variation. However, repeated low paste volume on pads beneath a BGA, inconsistent paste deposits on QFN thermal pads, or offset deposits near tightly spaced hidden joints are valid reasons to examine the post-reflow result by X-ray.

The same logic applies to AOI findings. A component may be slightly offset but still pass an AOI threshold because its body and markings remain within tolerance. Under an area-array package, that positional shift may reduce pad overlap enough to create weak or incomplete joints. X-ray helps determine whether the placement condition translated into an internal solder problem.

A useful escalation path is not “send every board with any alert to X-ray.” That creates unnecessary review load and can obscure urgent patterns. Instead, link the trigger to known failure mechanisms:

• Low or poorly centered paste deposits on hidden pads should trigger checks for incomplete joints or insufficient solder spread.
• Placement offset on area-array or bottom-terminated packages should trigger checks for ball or pad alignment.
• Repeated rework on the same package type should trigger checks for solder bridges, opens, pad damage indicators, or unusual void patterns.
• Thermal-profile changes, stencil changes, or new paste materials should trigger first-article X-ray verification on critical hidden joints.
• Unexpected electrical-test failures that localize to hidden-interconnect devices should be correlated with X-ray evidence before assigning a root cause.

This approach uses AOI SPI and X-Ray Inspection Equipment as a connected control system. Each stage supplies information that narrows the next inspection decision instead of operating as an isolated gate.

Void findings require a defined acceptance basis

Voids are one of the most frequent reasons that X-ray images receive attention, but a dark area in a solder joint is not automatically a reject condition. The relevance of voiding depends on the component function, joint location, heat-transfer requirement, current path, package design, and the acceptance rules established for the product and process. A small dispersed void pattern and a large concentrated void near a critical interface may carry very different implications.

Before using X-ray results for accept/reject decisions, the organization should define what the inspectors are expected to evaluate. This includes the region of interest, image orientation, magnification, thresholds or review criteria where applicable, and the treatment of borderline images. Without this preparation, different operators may interpret the same joint differently, leading either to unnecessary rework or to inconsistent escapes.

Void analysis should also avoid a common shortcut: treating void percentage as the only meaningful signal. Distribution matters. Voids clustered at an interface, evidence of solder starvation, irregular solder spread, or asymmetry across identical pads can indicate a process problem even when a single numerical measure appears acceptable. Conversely, a broad percentage value without context may not explain whether the condition affects function.

Use X-ray at the right point in the production flow

The timing of X-ray inspection depends on whether it is being used for process qualification, routine production control, troubleshooting, or final release. It does not have to be placed at one fixed location in every line.

During new-product introduction, stencil changes, package transitions, and reflow-profile development, X-ray often has the greatest value as a verification tool. It confirms whether hidden-joint behavior matches the process assumptions before volume production makes a defect pattern harder to contain. Inspecting representative assemblies at this stage can reveal whether paste apertures, placement settings, board finish, or thermal exposure require adjustment.

In stable production, full-board X-ray of every assembly may be appropriate only where criticality, package density, and cycle-time requirements support it. Other lines use targeted inspection of selected packages, first-off boards, defined sampling points, or assemblies that receive an upstream trigger. The correct choice should be based on the hidden-joint risk and the containment objective, not on a general assumption that more images always improve quality.

After rework, X-ray can be particularly useful for components whose joints cannot be confirmed from the surface. Rework introduces risks such as uneven solder collapse, shifted balls, trapped solder, or heat-related changes that are not visible after the component has been replaced. A post-rework image provides a direct record of the hidden area before the assembly moves on.

Interpreting images without creating false confidence

An X-ray image is evidence, not a complete diagnosis. A bridge may be clear in the image, but its underlying cause could be excessive paste, component movement, pad design, placement force, or a reflow condition. An open-looking feature may arise from a true joint separation, an imaging angle, component tilt, overlap from nearby structures, or insufficient contrast. Image review should therefore be tied to the assembly design and process history.

When an image indicates a defect, compare it with SPI deposits, placement records, AOI images, reflow conditions, and electrical-test information where available. This sequence avoids repeatedly treating the same symptom. For example, several adjacent BGA bridges may suggest a systematic paste-printing or collapse issue; a single distorted corner may point more strongly toward local placement, board warpage, or handling conditions.

Inspection programs also need protection against false calls. Excessively tight automated thresholds can generate review queues filled with acceptable variation. Loose thresholds can allow meaningful defects through. Build recipes using known-good assemblies and controlled review logic, then revisit them after meaningful changes in package, board finish, stencil design, paste, or reflow behavior. A program copied from a similar package is a starting point, not proof that the image criteria fit the new assembly.

Questions to settle before adding X-ray capacity

The equipment decision should follow the inspection requirement. Begin by identifying the smallest hidden feature that must be resolved, the number of joints or regions to inspect, the package mix, and whether the need is routine automated screening or detailed engineering analysis. Dense boards with overlapping structures may require different imaging capability than boards with isolated BGA locations.

Cycle time matters, but it should be considered together with handling and review time. An automated image capture sequence is only productive when fixtures locate boards repeatably, regions of interest are correctly programmed, and review rules are practical. Manual interpretation can be appropriate for engineering analysis and exception handling, yet it becomes a bottleneck when applied indiscriminately to high-volume output.

Traceability is another practical concern. For safety-sensitive or high-reliability assemblies, the inspection record may need to link the board identity, package location, image, program revision, disposition, and rework status. Decide early which images must be retained and how exceptions will be documented. Saving every image without a retrieval method adds storage; saving only a pass/fail flag may leave too little evidence for later investigation.

Signs that AOI alone is no longer enough

Consider adding targeted X-ray when hidden-joint components are increasing, when process changes cannot be validated from external appearance, or when electrical failures appear after boards have passed visual inspection. It is also warranted when a product’s function depends on heat transfer through underside pads, tightly controlled high-current joints, or concealed interconnects where a latent defect would be difficult to detect later.

The practical objective is not to replace SPI or AOI. SPI prevents defects by controlling the printed deposit before placement. AOI catches a broad range of visible errors quickly. X-ray closes the evidence gap beneath packages and other obscured structures. When inspection criteria, trigger rules, and image interpretation are defined around the actual joint risks, the three methods provide a more defensible release decision than any one of them can provide alone.