Messy. Messy. Messy.
Consider gluing these components, until you control your process.
It is unclear to us that a single factor drives these defects, but multiple factors that need to be tuned to work well together.
PAD DIMENSIONS * Some SM-782 pads are a bit archaic. 7525, Philips, and Topline pads are more in touch. Suppliers are the best source of information. Go there first. Everyone else is sheep. Baaaa. * At some point of development, the gap between pads is probably more critical that the size of the pads. Pads that are closer are better than those further apart. Something like 18 thou is better than the more common 25 thou.
REFLOW RECIPE: Have not tested different solder recipes, but be careful with fast ramp up.
SOLDER PASTE: Solder paste with same commercial characteristics ( Type 3, 63/37, No-clean ) gave different results.
IMBALANCES: Generally, tomb-stoning is caused by inequalities between the two pads, terminations, profile, etc. Search the fine SMTnet Archives for background. Beyond those general points, other things to consider are: * Evaluating effects of poor component termination plating. * Reviewing pad size relationships, since greater solder volume applied gets more tombstone results. * Reducing the purity of your N2. N2 increases the wetting action and decreases the surface tension of your solder, creating unbalanced forces on the component. * Checking placements not only for skew, but for exact alignment prior to reflow. * Assessing inner-wiring density of the board to determine if one side of the component may be getting hotter than the other, if this is a reoccurring problem. * Thermocoupling problem components and checking for a temperature differential, if all else fails.
FINALLY: I believe that with 0402, you have to be more concerned with placement accuracy than with other fly shit [well except of course 0201, ugh].
reply »