Has anyone had (successful) experience placing testpoint vias under BGAs? We already have conventional vias, but we're contemplating converting them into testpoints (probably by enlarging the bottom-side copper diameter to .036" and opening the bottom solder mask for a .003" clearance). We're planning to wave-solder the board, so we'll need to open up the topside mask, to eliminate out-gasing and/or trapped flux -- how much do we need to open the topside mask? Can we mask right to the barrel? I'm worried that opening the topside solder mask under the BGA will wick solder from the BGA land and/or compromise minimum clearaces between adjacent (unconnected) signals.
Can we keep the topside tented if we don't wave-solder?
Mark: The best routing of test vias for BGAs is to keep them far away from the device.
We route test pads for the outer two rows of balls and use vias as test pads for the inner rows of balls. Your bottom-side layout is fine.
Yes, on the top-side you can mask up to the via barrel with LPI. Barrel +0-0.003" is probably a good spec.
Here�s another thing to ponder as you�re grinding through this: your wave will probably reflow your BGA pads by conducting heat through those heat pipes next to each ball. So, you potentially have created a real mess for yourself.
Currently we slap a piece of Kapton tape over the via array before wave soldering. Consider plugging the vias with a conductive polymer. We are playing with some goop from DuPont. It would be nice if your designer could give you a thermal sink, but we know how that usually goes. It does give a good reason to have only essential test pads under the BGA!!!
Yes, you can keep the topside tented if you don't wave-solder.