Dye Penetration test for area array package failure analysis

First, what on the area array package are you analyzing � * Package cracking? * Cracked solder balls? * Er, what?

Second, most failure analysis discussions consider dye penetration to be a non-destructive test, but then again most failure analysis discussions center on engine blocks and similar chunks of metal, sometimes machined ceramics and other materials too. In our discussion here, it is almost definitely going to ruin the day of whatever your testing.

Third, failure analysis gurus prefer fluorescent dyes [Class I] over visual dyes [Class II], because fluorescences are produced in several sensitivities. But then again, those people are checking the carriages on amusement rides and we�re not.

Forth, responding your questions � Any recommendation on the dye to use? * Magnaflux [3624 West Lake Ave Glenview, Illinois 60025 847-657-5300 Fax 5388] Spotcheck Red Dye is a good general purpose dye for visual work. They also sell fluorescent dyes. * Other dye penetrant suppliers are: - Trikon Technologies 3375 Griffith St, St. Laurent Qc. (Canada) H4T 1W5 http://www.trikontech.com - Sherwin 5530 Borwick Ave Southgate, CA 90280 (562) 861-6324 FAX (562) 923-8370 http://www.sherwininc.com/

Any articles leading to this? * Contact the dye suppliers above. They know what they�re talking about and can focus on your specific issues. Most articles are aimed at the huddling masses of metal chip makers that we discussed earlier. * Here�s an overview http://www.amgas.com/penpage4.htm * General references are: http://www.ndt.net/abstract/wcndt96/pt_.htm * IPC has a standard test for checking for pin holes [http://www.ipc.org/html/2.1.2.pdf] using dye penetration that is a good starting point. [Don�t tell Steve that you downloaded the procedure. We don�t want him to feel bad that he purchased the whole thing.] * More specifically, "Reliability Of Ball Grid Array Packages In An Automotive Environment" Per-Erik Tegehall, Proc. Surface Mount International, 1997, pp. 85-92 analyzed ball cracking in BGA packages with dye penetrant methods.

Any comments on the effectiveness of this test? * It is very effective for locating gross leaks [10^-1 to 10^-5 atm-cm^3/sec] in electronic packaging. Recognize that after running the leak test, the package is ruined and further failure analysis is pretty much ruled out. * It is ineffective for locating fine leaks [10^-6 to 10^-10 atm-cm^3/sec] in electronic packaging. * In checking for solder cracking, it shows how much the dye penetrated cracks in balls and how much the dye did not penetrate. Before getting to that point though, area array packages are difficult to remove from substrates consistently and so to not damage things. - Per-Erik [above] glued a hook to the BGA package and hung weights from the hook to break the array loose after a couple of hours in a 85�C chamber. - Generally, we prefer a screwdriver or a beer can opener, but sometimes we make a real mess of some packages and / or substrates with these tools. [But then again, screwdriver is very good at removing screws. And the beer can opener � BINGO!!! You guessed it!!!] - Recently, I read an article [but, I can�t find it, don�t know where I saw it, wonder if I made it up, etc.] where someone was checking the cracking of solder balls in array packaging with dye penetrants. In removing the package, they were pretty unscientific, but not as unscientific as someone rooting around with a screwdriver. They [as I recall with my flux fume idled mind] secured the substrate with a clamp of some sort [or may be they cut the package from the substrate] and then used something like pliers or vice grips to wiggle the package back and forth, changing sides of the component periodically. Next time, we�ll think about this method, if we can�t find our usual tools of ignorance.

reply »