Bev,  Our investigation into gold embrittlement with SAC started with parts that are wirebonded internally after several other heated process steps.  In order to have bondable gold at the end of those steps, there is a minimum of 30 uinches of gold required inside.  Of course with plating processes, the outside gets thicker plating, so the spec for these was 30 - 80 uinches (80!).  One of the original thoughts was that we would have to remove the gold by double-dip.  (Messy, expensive, adds to the supply chain and leadtime -- ick)  The first trials were lumpy plus damaged the parts.  We put that aside to work with the suppliers on their gold spec, but I figure the issue may come up again, and I'd like to hear if it's been tried or is possible.

Dave,  I do like that paper in JEM, authored by our own Mike Powers.  I'm not sure how small the joint would have to be for ENIG to be an issue -- especially with the Au thickness recommendation dropping.  I doubt this will be a problem.  I posted a Au calculation spreadsheet on the TechNet site a while back.  Could run through the numbers to see.