News

If an optimized process can't meet quality requirements, reconsider the design.

Pb-free joints, on the whole, do not look as nice and shiny as SnPb. We are told that this cosmetic difference is not a defect. Then we also discover that many through-hole joints will not fill to the top with solder anymore. As a result, customers ask how we can optimize the process so that common quality criteria will once again be met. Unfortunately, it is not possible to meet the needs of the customer, even when the process is optimized for Pb-free soldering. SnPb soldering quality and appearance are related to solderability, while the solderability requirements for Pb-free are not the same.

Industry solderability requirements and test methods are based on SnPb. All test requirements - surface solderability, thermal solderability and specific soldering distance - are based on that solder. The test to check if components can withstand the soldering process without damage or deterioration is based on those alloys and their process temperatures of 245° to 250°C. Pb-free solders have a much higher melting point. SAC alloys have a melting point of 217°C, while Sn100C melts at 221°C, considerably higher than the 183° to 189°C melting range for SnPb solder. The higher temperatures will have a major impact on solderability - surface and thermal.

On boards with PTHs, solder can only wick up into holes or joints through capillary action. This works only when the solder is liquid. Thus the solder on the top-side of the joint must have a temperature of at least the melting point of the solder to create a sound joint. For Pb-free solder, this temperature requirement is 28° to 38°C higher. The design of the joint must enable the process to fulfill the demand for good joint quality.

If we cannot meet the quality demands with an optimized process, the design must be reconsidered.

When new solder alloys are used, many components and board layouts will not pass the solderability test requirements for SnPb solder. These boards and components may not meet new process requirements. Higher preheating and solder wave temperatures and longer dwell times may be necessary. All components and the board must be able to withstand this higher thermal load, and fluxes must be able to function well with these demanding settings.

Conversely, the joint design must be such that the solder will not solidify during filling of the joint as a result of the heat sink effect of the board layout or the component lead shape. Insufficient surface solderability or insufficient flux action can result in insufficient hole fill.

I would assert that industry has not yet looked in detail at what the new solderability requirements for Pb-free soldering should be, or how they should be defined. In principle, the same test methods used for SnPb solder seem to fit. However, the test temperatures and test times should be reconsidered and redefined. This will result in new soldering distances for each component. Also, thermal issues related to board design need to be investigated. The "homework" done decades ago in laboratories for SnPb soldering should be revisited, and not only for the Pb-free processes and its melting temperatures. This must be done not only for the process, but also for the suitability of components to higher temps.

New test requirements should be worked out for boards and for the board/component combination that forms the joint design. Finally, joint design should meet solderability requirements so that reliable, sound joints can be repeatably formed.

If the joint design is not optimized for Pb-free, one may get unsatisfactory results that point to an unstable process. In a sound design, small process variations will have no effect.

Changing to another flux in the hope of improving solderability is often not a viable option.

Looking at the surrounding components proves that such effects are often related to specific joints and not to the overall process, particularly if the surrounding joints are soldered well. If specific joints are not meeting the quality demands with this process, there must be something wrong with them. If the process were not sound, all joints would show the same effects or defects. If the majority of the joints are good, look for other issues related to the final result of that process for a particular joint.

To learn if such effects can be permitted, investigate whether the reliability of such joints will be sufficient to meet the demands of the product. If so, one can make allowances for the fact that these joints do not meet the visual quality requirements, as the reason for the effect is known. If the effect is classified as a defect, the joint(s) involved are thus unsuitable for that process, because one is unable to compensate for the variables causing this defect. Redesign is often needed.

Gert Schouten is a senior engineer at Vitronics Soltec (vitronicssoltec.com); gschouten@vsww.com.