I am running a product with board size ~18"X16". I faced stencil aperture positional accuracy issue which slightly shifted at both end of stencil at positive and negative in X direction thus causing solder paste to pad mis-registration. Due to this solder paste to pad mis-registration, it created several SMT defect as such misplaced, tombstoning and missing SMT component. The question that I would like to consult are: (1)Is this condition normal for board size greater than 15"? As I have observed several stencil that we have on hand at almost similar dimension facing the same issue. (2) What is the short term and long term action to eliminate this type of SMT defect? Thanks.
I haven't been in a bad enough situation to try it yet, but I had a conversation with Alpha Metals when we started doing 0.4 mm pitch and they mentioned that if we got into a situation like yours we could send them a board and they'd modify the artwork to create a stencil that accounted for stretch so that apertures would line up with pads on both ends of the board.
We have boards larger than 15" but we're in the high reliability market so we tend to use high quality (and expensive) board suppliers which I assume control stretch a little better than others.
We work with a lot of larger boards over 24", haven't run into this issue yet. I agree with scott could be your board supplier if they are lower quality, from something as simple as unit conversions in CAD to dealing with different temperatures. Also could be your screen in cutting temperature and mounting tension. These effects are unnoticeable at smaller distances but apparent across an 0201 or 0.4mm pitch IC. Call your stencil provider see if can work with your sample, if its consistent and uniform they can simply add a stretch coefficient while cutting.
Our lasers are accurate to +/- approx. 5 microns over 20 "X-Y. I have seen boards 18 x 24 stretch 7mils. I would agree with others, the only way is to send a board to you stencil vendor so they can edit the file to match the board. This is very common today. Ultra fine pitch components do not give you the luxury of misaligment or misplacement. We have 9 facilities that could help you if required.
I have fabricated a artwork of stencil on a mylar and make comparison on PCB and Stencil in order to confirm if there is any PCB stretching at both end. The results are comparable for both stencil and PCB without any stretching . So,I assumed that there is no PCB stretching for this case. However during printing, we still see printing offsetted for 2-3 mils shift at both end of PCB after get a very good board support during printing. Do you still see any other factor that will contribute to this failure? Btw, I have modify the opening of stencil for 0402 pad from home shape to round shape (solder volume increase for another 20%, the result is promising which yield improve from ~60% to ~90%.