Is your selective solder process causing leakage failures?

For years clients have brought us a variety of ill-performing electronics hardware. We see a lot of interesting stuff like tin whiskers, open resistors, dendrite shorts under conformal coating, and a variety of burned assemblies (even with just a 3.3V battery power supply). These failure mechanisms have residue from the process/fabrication or outside source that, in the right location, is a critical variable and causes failures. Such failures leave behind a visible element where the damage can be observed; the element is seen in just one or two locations on an assembly, not over the entire surface or on all components. Localized extraction techniques have been used for years to isolate and compare problem areas vs. areas that did not show problems in order to understand chemical differences in ionic and organic residues.

These techniques are used in a growing number of cases where no visible degradation has occurred, nor innerlayer issues seen, but there is measurable stray voltage (electrical leakage) around, across and below components or trace-to-trace over the soldermask. This stray voltage phenomenon has a greater effect on the surface of the assemblies on high impedance circuits. In historical work, we have seen issues stemming from component cleanliness, wave-solder fluxing, HASL bare board flux, and overspray from a no-clean fluxing system – all responsible for leakage and stray voltage poor performance. Now we see a greater occurrence of leakage failures due to selective solder pallets over a standard wave solder system or a robotic selective solder system, as outlined in this case study.

An electronics assembly was presented for investigation from a no-clean process using double-pass reflow and selective solder pallets over the wave. The performance problems were in moderate humidity at 65% RH; the high impedance areas of the circuit would cause the system to reset or shut down amid normal operations. This was seen in the field and repeated in the environmental chamber on the failed assembly. If the assembly was baked and kept in low humidity conditions, the problem did not present itself. Only after exposure to 65% RH did the problem reoccur. This was also found to be true on most of the current production samples (nine of 10) evaluated under environmental exposure testing. Only one of 10 assemblies did not exhibit poor performance in high humidity conditions.

Using ion chromatography and localized extraction, we looked at five assemblies in three locations each. Production cleanliness data records indicated that all tested assemblies for the current production and field return date codes passed the 10 µg/in² of NaCl equivalence. These assemblies were 8 x 4" with a dense SMT component population.

Using localized extraction was the crystal on the bottom side near the selective pallet area (which is part of the high impendence circuit), the selective solder connector, and a topside via area connected to the circuit.


IC results show the selective solder area had acceptable levels of flux that are heat-activated and meet Foresite-recommended limits. The area of the crystal near the selective solder process for the two failures shows high WOA only, indicating a heavy fluxing and trapped, unreacted flux on the crystal and on the topside of the via near the selective solder area. This level of WOA from the wave soldering flux is moisture absorbing and conductive and causing the leakage issues. Compared to the current production board that passed humidity testing, the level of WOA shows a greatly reduced level of flux compared to all the other current production. This variation is currently under investigation.


Using localized extractions with IC, we are able to see the ionic differences from soldering and fluxing processes that are creating stray voltage areas through the partially heat-activated no-clean flux residues. These residues are moisture-absorbing and in most cases insufficiently corrosive to cause a dendrite short, yet able to create a leakage path capable of a measurable short. This leakage path can drain batteries, influence critical readings and shut down high impedance sensitive circuits. This investigation shows no-clean, low-solids flux residues that are not heat-activated will cause performance problems, but must be investigated by localized extraction techniques to isolate the small pockets of contamination that cause the problems.

Terry Munson is with Foresite Inc. (residues.com); tm_foresite@residues.com. His column appears monthly.

• Prev
• Next