First, most �thermal shock� recipes are not a reliability test. There is no relationship between the failures seem in these tests and in-use application. These tests show you the failures in these tests. That�s it!!!
Second, accelerated life tests, according to SM-785 - Guidelines For Accelerated Reliability Testing Of SM Solder Attachments, is a more realistic approach to solder connection life testing than thermal cycling.
Next, on a yellow flux residue that became darker during the thermal cycling: Yup, this can happen.
On transparent flux residues that became white during the thermal cycling: This was probably caused by water condensing on your board during the cycling.
On applying silicon coating over your low residue flux: If it works, it�s good. We see some coatings that work with some flux residues and some that don�t.
Any other test? * Compare the ionic contamination levels of your old flux and new flux. * Run surface insulation resistance tests on your new flux.
Continuing, J-STD-001, Appendix B talks to the qualification of new processes and process materials.
reply »