During BGA rework verification, we found that the same chip experienced functional failures across three consecutive batches. After desoldering, visual inspection showed completely normal appearance, and open circuits were only reported one after another when running aging tests on the machine. Later, when using the JIAJINYUAN X-Ray inspection equipment set to the 80kV/120μA parameter range for scanning, we found that pin 7 in column 4 had typical cold solder joint, with a continuous gap between the solder ball body and the pad, and the adjacent solder ball at this position also had slight solder ball offset, with an offset of about 32μm, which has exceeded the upper limit of the ±25μm tolerance allowed by IPC-A-610G for Class III products.

Solder joint voiding is particularly stubborn on large-size copper substrates. For example, a certain power module uses a 6-layer thick copper PCB paired with a 0.8mm thick heat dissipation copper block. After reflow, X-Ray inspection shows that the void rate of multiple solder joints in the central area is concentrated between 35% and 48%, far exceeding the customer's required red line of ≤25%. Our investigation found that this is not due to insufficient Solder Paste activity, but the stencil aperture ratio is set to 0.78, resulting in small printing volume. In addition, the reflow oven chain speed is too fast, causing insufficient volatilization of the Flux, and residual gas is trapped inside the molten solder and cannot escape, eventually forming regular circular voids during the cooling and solidification stage.

Solder ball offset on 0.4mm pitch QFN devices is easily misjudged as tombstoning. In fact, as long as the JIAJINYUAN X-Ray inspection image is switched to the transmission overlay mode, it can be clearly seen that the bottom of the offset solder ball still maintains a metal connection with the pad, while for a truly tombstoned device, the solder ball will be completely lifted off the pad surface. This difference is difficult to distinguish with the naked eye at 200x magnification, but the grayscale contrast of X-Ray inspection is sufficient to distinguish the physical contact state between the two.

We once encountered a batch of medical imaging boards where X-Ray inspection found a large number of tiny solder beads scattered around the BGA solder joints, with diameters concentrated between 8 and 12μm. At first, we suspected that the stencil was not cleaned properly, but later traced it to fluctuations in the vacuum value of the mounter nozzle, which caused slight tremors during chip placement, causing the edge of the solder paste to be squeezed and splashed. This type of defect is completely invisible under AOI, and only the high penetration of X-Ray inspection can capture the shadow of solder beads deposited under the solder mask.

For 0201 and smaller size components, X-Ray inspection must be paired with the unique focus drift compensation algorithm of JIAJINYUAN equipment. Otherwise, image blur caused by focal length changes will distort the edge contour of the 0.15mm pad, and further misjudge the normal solder joint wetting angle as a void boundary. We have tested that when this algorithm is not enabled, the false positive rate for void identification of 0.1mm pitch solder joints reaches 17%, and after enabling it, it drops to within 0.8%.

In mass production, X-Ray inspection cannot only look at single-frame images, and must continuously collect more than 3 frames of dynamic sequences. Because some cold solder joints will experience transient conduction during thermal expansion and contraction, static images are prone to missed judgments. We once found in the inspection of a batch of automotive T-Box motherboards that pin 11 showed good connection in the second frame image, but in the third frame, internal fracture cracks were exposed due to solder joint deformation caused by slight vibration of the stage. This kind of dynamic abnormality can only be captured by the 120fps high-speed acquisition capability of JIAJINYUAN equipment.