Issue link: https://iconnect007.uberflip.com/i/1546933
52 SMT007 MAGAZINE I OCTOBER 2026 Even when the solder is completely removed, light still reflects more brightly off the copper or gold pads beneath it. Applying a coat of white paint maintains consistent surface brightness and miti- gates noise due to damage caused by the shearing process. However, ball removal alters the sample composition, potentially affecting warpage and thermal mass. Alternatively, paint can be applied to the surface while the solder balls are still intact. Post processing requires digitally selecting a region of pixels that encompasses the solder ball and scanning the entire image looking for matching areas. Processing time is dependent on the number of pixels for both the solder ball region and the full image. Searching for a 14 x 14 pixel area in a 750 x 750 image will take a greater amount of time than searching for a 14 x 14 pixel area in a 500 x 500 image. Solder balls must be removed from all acquisitions taken throughout the thermal profile. This can be accomplished either by repeating the previously described process for each individual image or by masking multiple images using the ball location from a single refer- ence image. As samples expand, contract, or shift during the thermal profile, the masked regions may become misaligned. Careful attention is required to ensure the correct areas are masked for each temperature. The methods presented in this article eliminate the need for solder ball shearing and painting. However, the primary advantage to these methods is the simplified data processing and time savings associated with it. DIgital Fringe Projection Digital fringe projection (DFP) is an optical metrology technique that measures surface contours by projecting fringes onto a sample with a digital projector and observing the resulting fringe distor- tion. Unlike SM, the fringes in DFP do not directly indicate height changes. Instead, the height varia- tion is determined by comparing the distortion of measured fringes to that of a reference plane. The fringe change relative to that of the reference plane can be determined by subtracting the refer- ence plane phase map from the sample surface phase map. DFP has minimal limitations in terms of data density, allowing smaller pixel sizes. This is advantageous to show more surface detail or measure small features of a sample surface. DFP is better suited to resolving fine surface features on smaller samples, such as solder balls. However, the measurement resolution gets worse as FOV (field of view) increases Devices Under Test This study used four different samples, all smaller and thinner BGA style devices, aligning with the focus of this study. Each BGA is assigned a number, with sample details provided in Table 1. A deeper analysis was conducted on the BGA 1 samples to further refine the correlation between top and bottom warpage. Table 1: Dimensions, ball diameter, and ball pitch of tested BGAs