Issue link: https://iconnect007.uberflip.com/i/1546933
36 SMT007 MAGAZINE I OCTOBER 2026 and documented. The stencil was cleaned with a wet/dry/vacuum cleaning profile after each board was printed. The PCBs were processed through pick-and-place and then reflowed. Each board was inspected for voids using DAGE Quadra 7 X-ray inspection. The voiding data was then analyzed with JMP® statistical software. Reflow and Inspection Methods An eight-zone convection oven was used for the reflow process. In this investigation, two addi- tional profiles were introduced to optimize key reflow parameters, including soaking time, peak temperature, time above liquidus, and heating ramp rate. The primary difference between the two new profiles was the ramp rate from 180°C to 220°C, set to 1.0°C/s and 1.5°C/s, respectively. For temperature measurement, one thermocouple was placed under each QFN location (QFN 1, QFN 5, QFN 8, and QFN 12) to ensure accurate readings of the test vehicle PCB. An addi- tional thermocouple was used to monitor air temperature, resulting in a total of five thermocouples. Voiding Inspection After reflow, the PCBs were inspected for voids with the DAGE Quadra 7. The voiding data captured was then analyzed with JMP®. The goal was to find the best combination of solder paste, reflow profile, and stencil thickness, and board metalliza- tion to achieve an overall average voiding amount below 15%. Analysis Across the three trials, a clear progression in voiding performance was observed, shaped by the combined effects of surface finish, stencil thickness, alloy, and profile, as shown in the one-way ANOVA overall voiding and largest void area analyses in Figure 2. Trial 1 (OSP, 4 mil) showed only a modest impact from reflow profile modification, with overall voiding for Paste 2 reduced to 11% and 9.0%. Trial 2 (ImSn, 4 mil) followed a similar profile-driven trend but with consistently higher voiding performance, with Paste 2 reduced to 14.3% and 13.1%. The higher voiding observed with ImSn underscores the importance of surface finish, particu- larly in combination with SAC305, which exhibited the worst performance in this study. Figure 2: One-way ANOVA overall void data analysis and Tukey-Kramer comparison for Jabil boards with QFN components for all trials.