We recently got our last BGA X-Ray results back (no X-ray in house yet) and had more than one board/BGA look like the attached image. The balls on one corner are completely collapsed (so much so that bridges appear) and raised on the opposite corner. Pitch is 1mm
This is a very low-volume board. Last run was 4-5 months ago, and out of 15 boards there were no issues. Can't think of anything that would have changed in our process to suddenly create this problem.
Other BGA's on the board were fine, just this particular part. The board contains qty (4) of this problematic BGA, and the problem appeared in 3 different locations. At one location, (3) different boards had problems at the same corners (upper left). At a different location one BGA had the upper-left over-collapsed and on another board it was the lower-right corner that over-collapsed.
Half of the boards had this problem, the other half were fine. We're trying to figure out what went wrong, but since there's so much time between the production run and the x-ray results we're kind of shooting in the dark. Any guesses? Placement issue? Reflow? Solder print?
Two thoughts from me: 1)Are you sure the balls are collapsed? Have you looked at the perimeter from the side with a scope? From your x-ray, it appears you are correct in saying they are collapsed as the balls get smaller in a uniform fashion as you move away from the bad corner but it is 2D. I would guess that defect could also be caused from excess solder paste if the PCB is not flat to the stencil or paste was forced under the perimeter of the apertures. 2) Have you performed a profile using a mole? It could be that you are on the very upper limits of TAL and the balls are on the verge of collapsing and it is occuring in an area where the PCB holds heat for longer or conversly heats up more easily.
Thanks for the input! In looking at the image again, along w/ some of the others we were supplied with it does look like it could just be too much solder. As mentioned, that does look like a lot of solder bridging up there.
It will likely be a while before we're running this again, but I can post results if anyone is interested.
Check your reflow profile as well, it could be that your ramp rate from flux activation to reflow might be too steep. This is in the same line as the post from Ken. Too fast a rise in temperature will collapse the outside rows too far in advance of the inner rows, leading to that kind of corner defect.