CUSTOMER ISSUE: The caller was observing solder joint separations within a surface mounted transformer. They stated that initially good solder wetting was observed after re-flow of the paste. However, soon after, the component was able to be lifted off of the board. The customer provided ten pristine components, along with an assembly with obvious cracks, referred to as the "bad" assembly. They also provided a "good" assembly with no apparent separation
by Sam Pepe
Investigation Technique:
The customer requested quantitative solderability testing on the transformer. Since the component was a leaded surface mounted component, the solderability of the transformer was evaluated in accordance with IPC J-STD-002B 4.3.1 test E - Wetting Balance Test (Leaded Components). Since the failure was occurring with this particular component the source of the failure was thought due to the component and not the board. In addition to quantitative solderability testing, the following analysis techniques based upon ACI procedures were used to evaluate the failure: optical microscopy, micro-sectioning in conjunction with scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS)

The component incorporated four pins/leads which made direct contact with the pad areas. As a result of this configuration, it was necessary to section the leads from component. Only five of the ten components were analyzed for a total of ten actual measurements. Based upon the information provided to ACI by the customer, the leads were composed of a copper substrate with ductile nickel which had an electrodeposited tin finish 4 x 10-4 inches thick

Both the "bad" and "good" assemblies were prepared for micro-sectioning during which measures were taken not to stress the solder joints. However, because of the weak solder joint and the vibrations generated from the cutting mechanism the component on the "bad" assembly dislodged from the board surface. It was determined that the sectioned board area from the "bad" assembly be prepped for SEM analysis without further treatment except for sputter coating of gold (Figures 7 and 8).

Results:
Visual Inspection of "Bad" assembly before destructive testing:

Wetting Balance:
Initial analysis of the base of the leads was not able to generate a detectable wetting force as a result only visual analysis was pursued (See Figures 3 through 6 which show the base of two leads before and after soldering).

In summary, two of the ten leads analyzed showed poor solderability similar to Figure 4. The other eight leads tested showed partial wetting as in Figure 6.

SEM analysis of "bad" and "good" assembly: All four pads had similar appearances as that in Figures 7 and 8.

Conclusions:
The components did not generate a measurable wetting force; they did not comply with J-STD-002 B quantitative solderability standards. Most leads tested exhibited de-wetting. The poor wetting or de-wetting is probably the result of the inconsistent surface plating. In some instances voids as in Figure 11 were observed allowing the underlying nickel surface prone to oxidation making the surface un-solder able.
The large voids observed at the pad of the dislodged component (Figures 7 & 8) are reminiscent of volatile gases trapped in the molten solder during re-flow and can be the result of excessive flux, improper re-flow profile, poor flux selection or poor solderability.

The poor wetting from the exposed nickel along with the voiding at the pad surface resulted in a mechanically weak solder joint.

Recommendations:
Confirmation of a uniform, proper plating thickness was recommended in addition to routine Wetting Balance analysis of incoming components. Pursuit of a second source of transformers was also recommended. Such qualification and consultation support would be available through ACI.