Issue link: https://iconnect007.uberflip.com/i/1389320
12 SMT007 MAGAZINE I JULY 2021 Rhino Team from IPC really gives a lot of power back to the assembler: "If this is my end-use environment, I should probably do this kind of testing." Now, it's more of a guideline: "is is how you come to the conclusion that your boards are clean enough." It's the old ques- tion nobody can answer: How clean is clean enough? It depends on your end-use environment, and people are coming to accept that they must stop relying on outside organizations to determine how clean their boards need to be. Now the IPC is backing up that thought by saying, "You need to figure out how clean your boards are." And I think that was the impetus for completely changing sec- tion 8 of J-STD-001. Johnson: If perceptions and requirements for cleaning are changing, and if we're getting into a period where cleanliness is becoming increasingly important, what are the drivers for the assemblers? Camden: e drivers are that they've recom- mended objective evidence tests. With surface insulation resistance, it will be a prerequisite for any new product going forward for accep- tance (if you're looking at IPC). Surface insu- lation resistance was really a test that's been agreed upon, and rightfully so. ere are a cou- ple of different ways to do it. ere are ways to tweak that test and still stay within the spirit of the IPC TM-650 test methods. But they recom- mend the SIR, and then you can build on that with your own ROSE testing. If you have an acceptable lot of boards through SIR, then you take a lot of boards built with the same materi- als in the same time frame and do ROSE test- ing to determine what that number is, based on their equipment. what to put in place instead of accepting that historical number. Cleanliness and reli- ability go hand in hand, and I don't see it changing any time soon. And when you put miniaturization on top of that you're just compound- ing your own problems. Johnson: You were mention- ing work with IPC standards to change ROSE testing and so forth. From your perspec- tive, what's going on there? Camden: When you think about it on the sur- face, it seems pretty simple. e ROSE tes- ter, developed by the DoD in the 1970s, was never meant to be acceptance criteria. And when you think about the material sets and the components that were being used at that time, when this go/no-go number was put together and generally accepted by a lot of the manufac- turing sector—the boards were full rosin, they were being cleaned. If they were being cleaned, they were being cleaned with Freon. None of the technology has really survived as-is from the 1970s, when that criterion came out. Over time, regarding miniaturization, as we see less ionics being allowed on the board to function properly in the field, there had to be a better way to determine what test was required for your particular board. e industry is start- ing to realize that all boards are not made the same. We've all known that, but we've got the failure data to tell everybody that your board is very unique. Once they decided that 1.56 num- ber really was no longer applicable, then it was just taken out. A lot of companies were mov- ing in that direction on their own, as they saw their le hand was holding a report that said they had met certain criteria, and in the right hand was their failure analysis report. When you see enough of those, you can conclude that maybe this isn't working. at Eric Camden