Hi all,
Long story ahead.... We are having issues with board vendors on some of our higher layer count (thick) boards. Specifically under BGAs with high aspect ratio vias (14:1). We are using ENIG finish and a lead free process.
Basically we get two or three (sometimes many more) random pads under the BGAs where the gold is still perfectly in tact after reflow and removal of the BGA. It also occurs on decoupling caps under the BGA. We've verified our profiles. I've also done simulations without a component by applying solder paste to the pad and using hot air to reflow the solder. The pad does not wet, the gold does not flash off. However if I use and iron and lightly move the iron over the bad while soldering I can get the pad appear to wet. I cannot however confirm an intermetalic. (no SEM in house!) Failure rates on affected boards range from 10% to 40% failures.
These defects only occur in areas of high density, high aspect ratio vias. It can occur on either side of the board (BGA side or decoupling caps). We know there's something over the gold but we can't be sure what.
We've had the issue with three different board shops, each has come to their own "solution".
1. The first shop has changed it's rinse/dry cycle on these cards. The theory is that there is chemistry trapped in the vias that is affecting solderability. We've only ran limited boards from this shop since the issue. It took them so long to figure it out that we had to move the boards to other shops.
2. The second shop moved to a SMOIG process. Keeping the vias clear of SM to help the cleaning process after ENIG. So far we are seeing better results but recently we've hit a small patch of issues but these appear to be more like black pad not the exposed gold issue.
3. The third shop is struggling. What's odd is they were our best shop and have been building these difficult boards for us for almost 2 years with no issues. Now we're seeing this creep in on several PNs. Their initial root cause was incorrect mesh size used during via tenting. They implemented changes but now we've seen it again. We've tried to hint at the solutions from the other board shops but they have been avoiding looking in that direction.
In the end I'm not sure any of the shops have really found the root cause.
I'm sure I've missed some key info along the way but that's the basics. If you have any ideas on what the real root cause of these issue could be please comment.
Thanks
reply »