Analysis of solderability test methods: predicition model generation for through-hole components

In order to achieve a reduction in solderability related defects on electronic components and Printed Circuit Board’s (PCB’s) in electronics manufacturing, preventive controls such as “Dip & Look” and “Wetting Balance” solderability testing need to be fully optimised to screen out all poor soldering components and PCB’s. Components and PCB’s that pass these tests should solder correctly in volume production. This thesis initially investigates the variations and effects of the Dip & Look solderability test on components and PCB’s. Data from this analysis proved that no matter how extreme the oxidisation on the component termination or PCB pad, the visual inspection criteria of 95% solder coverage is achieved each time. Dip & Look testing therefore serves no useful purpose to the electronics mass manufacturer in determining the solderability of a component or PCB. The second option available to screen components is the Wetting Balance Test. Due to the variation of parameter settings within the international standards for solder temperature, immersion speed, immersion depth, removal speed and dwell time, a complete analysis was required to determine the optimum settings for the MUST II Wetting Balance machine that would detect poor soldering components. The test specifications vary considerably between all the international standards. Design of Experiments conducted an in-depth analysis to determine the optimum Wetting Balance test settings using the range of test specifications stated within the standards. Within the range of specifications the least stringent and most stringent settings were developed which highlighted the difference in results when testing at the lower end and higher end of the current international standards. Prediction models were generated for each of the responses (Ta, Tb, Fmax, TFmax, F1 and F2) using Wetting Balance machine parameters solder temperature, immersion speed and immersion depth. To test these models, components with a history of solderability issues were tested and evaluated using the least stringent and most stringent settings. XRF measurements were conducted to ensure uniform plating thickness. Both components passed the Wetting Balance test criteria generated by the model equations using the least stringent settings but failed when using the most stringent settings. The current industrial specification for Ta (<0.6 seconds) and Tb (<1 second) were also achieved on both components even though the components had known soldering issues on a series production line. It was proven that there was a 40% difference in the Ta and Tb values when testing at the lower end of the international specification as opposed to testing at the higher end of the specification, which questioned the spread on the tolerance of the parameter settings within the current international standards. It was established through experimentation that the current F1 criteria, which states no less than 50% of Fmax, must be reviewed based on the analysis carried out in this thesis. To ensure completeness a component with no soldering issues was also tested using the same procedure. This component passed the Wetting Balance test using the least stringent and most stringent settings illustrating that the settings derived through this research are robust to detect good and poor soldering components. This research has developed an alternative set of Wetting Balance test specifications and has defined new model equations that will predict the Wetting Balance responses such as Ta, Tb, Fmax, TFmax, F1 and F2, which will result in components which are deemed acceptable under international and industrial standards.