Lynda Pelley, Teledyne DALSA Inc., Waterloo, ON, Canada
Hugo Gallegos, SMTC Inc., Chihuahua, Mexico
New product introduction at an Original Equipment Manufacturer (OEM) typically includes reliability testing in the form of thermal cycling. Samplings from the engineering models of a new product line were consistently failing the optical testing following the reliability testing. Cross sectioning revealed stress fractures in the solder joints of the LGA ceramic image sensor component. As the image sensor dominates one side of the PCA, the root cause appeared to be Coefficient of Thermal Expansion (CTE) mismatch between the FR4 material of the PCA and the ceramic body of the image sensor. As this new product line was to be high volume, the goal was to avoid introducing an additional component under fill process step and resolve reliability issues through basic SMT assembly processes at the Contract Manufacturer (CM).
A joint OEM-CM Tiger Team explored the contributions of both the solder paste volume and the reflow temperature profile to the robustness of the LGA solder joints. The stencil design aspect included the shape of the aperture and the resulting volume of solder paste deposited. The solder paste volume affects the standoff of the component from the PCA. The sensor manufacturer datasheet for the oven reflow profile allowed a range of temperature values, the most significant being the temperature cool down slope. The cool down slope affects the granularity of the solder joint. The CM’s team performed a DOE maximizing and minimizing both the solder paste volume and the temperature profile cool down slope combinations.
The DOE results trended towards higher paste volume and steeper temperature cool down slope. The risk of solder bridging using the higher paste volume was minimal, but the steeper temperature cool down slope produced unacceptable process voids. Therefore, the choice for optimal SMT process condition was a combination of high solder volume and lower temperature cool down slope. Passing results for optical testing post reliability testing, followed by cross section, validated this optimization.
Based on the DOE, the stencil aperture design has more influence on overcoming CTE mismatch between the ceramic LGA and the FR4 PCB than the reflow oven temperature profile cool down slope. Considering the results of the DOE and the identified risks with each approach, the overprinted square stencil aperture and the lower temperature cool down slope 1.8°C/second was chosen for the process design. The NPI was successful. Post DOE and engineering trials, there have not been any further CTE mismatch issues with the regular production of this product.
Initially Published in the SMTA Proceedings