When we encounter large-area voids accompanied by the pillow effect on BGA solder joints under X Ray on the production line, we must simultaneously check the viscosity change and metal content stability of the Solder Paste batch, and confirm whether the release rate of this batch of solder paste has decreased under the same stencil aperture, because uneven solder paste printing thickness will directly cause the solder ball to fail to fully collapse during reflow and form a metallurgical bond with the pad.

Improper matching of stencil aperture size and thickness is a high-frequency cause of solder paste defects. For example, when a 0.12mm thick stencil is paired with an 80μm circular aperture and the solder paste particle size distribution is too wide, a phenomenon of local solder volume redundancy coexisting with Flux residue will form at the bottom of the BGA. Such residue is more likely to hinder the wetting and spreading of molten solder on the pad in a nitrogen reflow environment, ultimately presenting as ring-shaped voids with clear edges and dense centers under X Ray.

Improper combination of squeegee pressure and speed will cause shear tearing of the solder paste in the BGA pad area, especially on devices with a pitch below 0.4mm. If the squeegee angle is lower than 55 degrees and the stroke speed exceeds 50mm/s, microcracks will appear at the edge of the solder paste and be propped open by flux gas in the early stage of reflow, forming the unique "solder ball suspended above the pad" structure of the pillow effect. At this time, X Ray can not only see voids but also often a slender shadow band perpendicular to the pad.

If the heating slope of the second zone of the reflow oven exceeds 2.5℃/s, the rosin-based flux in the solder paste will boil violently and entrain bubbles into the molten solder. However, an oxygen content below 100ppm in a nitrogen environment inhibits bubble escape. These trapped bubbles solidify into regular circular voids after cooling, with diameters generally concentrated between 30 and 60μm, and mostly distributed at the center of the solder joint rather than the edge.

We once compared the same solder paste on two SPI devices and found that when the SPI calibration deviation reached ±2μm, the solder paste volume judgment error for 0.3mm pads was as high as 17%, which directly misled the subsequent setting of stencil compensation parameters. Therefore, before troubleshooting solder paste defects, we must first use a standard block to complete the zero reset and height calibration of the SPI optical system, otherwise all printing data will be untrustworthy.

If the thickness of the nickel layer on the BGA pad surface is lower than 150nm, local palladium layer exposure and oxidation are prone to occur after multiple high-temperature reflows. This oxide film will seriously hinder atomic diffusion during solder paste soldering. Even if no obvious voids are seen under X Ray, the solder joint thrust test will fall below 2.8kgf. Therefore, for high-reliability products, ENIG plating thickness and solderability must be sampled and inspected for each batch.