I recently ran across an interesting presentation on tin whiskers, by
Lyudmyla Panashchenko of NASA, at:
http://nepp.nasa.gov/whisker/reference/tech_papers/2012-Panashchenko-IPC-Art-of-Metal-Whisker-Appreciation.pdf
Apparently tin whiskers (and cadmium and zinc whiskers) aren't a
newly-discovered phenomenon, but were known back in the 1940s and
1950s. Indeed, that seems to be why lead has been added to tin
platings for all these years. And it only takes one or two percent of
lead to suppress whiskering. (And that in an already-thin tin
plating; in a world where lead-acid car batteries are made by the
millions, banning lead in tin platings seems pretty silly.)
Whiskers are hard to detect visually; besides a microscope you really
need the right lighting, as they are so thin that otherwise they won't
be visible. (The presentation goes into details, including video.)
Whiskers are also hard to detect electronically, since they burn out
so easily, and since they grow insulating films. These effects also
"can explain survival of some electronics in the field despite whisker
infestation". It commonly takes only a couple of hundred millivolts
to burn out a whisker. But when a whisker burns out it may start an
arc, especially at low air pressures or in vacuum. (Thus the
particular relevance of this topic to NASA and to amateur rocketry...)
Due to these difficulties in detection, the whisker problem is
commonly severely underestimated.
Conformal coatings can help, even though whiskers punch through some
coatings under some conditions: the emerged whisker can't repenetrate
the coating at another location, since it is thin and easily buckles.
They also help against whiskers that are floating around in the
environment (say, zinc whiskers from a nearby piece of galvanized
steel).
--
Norman Yarvin http://yarchive.net/blog