BGA ball size reliability

On 1.27mm (0.050 inches) Standard Ball Grid Array's. The ball size diameter starts at approximately 0.025 inches.

If the ball size decreases to 0.012 inches (due to the solder flowing down a via) is the solder connection reliable?

This is a known Liquid Photo Imageable (LPI) Soldermask clearance violation. I have checked numerous seminar notes I have been to and a couple of hard bound technical books and came up empty as to any discussion pertaining to ball size and reliability.

I am under the impression that if we have less than 50% of the ball diameter remaining that it would be a defect.

Thanks

reply »

| On 1.27mm (0.050 inches) Standard Ball Grid Array's. | The ball size diameter starts at approximately 0.025 inches. | | If the ball size decreases to 0.012 inches (due to the solder flowing down a via) is the solder connection reliable? | | This is a known Liquid Photo Imageable (LPI) Soldermask clearance violation. I have checked numerous seminar notes I have been to and a couple of hard bound technical books and came up empty as to any discussion pertaining to ball size and reliability. | | I am under the impression that if we have less than 50% of the ball diameter remaining that it would be a defect. | | Thanks | | | We allow no solder ball size reduction. We allow no solder paste/ball drain off into any hole type. We paste during production, of course, and rework.

This may change when we have objectively determined certain micro via structures acceptable. This will only allow minimal solder ball size reduction. We consider reduction to be one of two things - either partial reflow (viewed with X-Ray as insufficient solder) or excess solder when balls are very close to bridging (a little more complicated but a function of too large pad sizes [design violations] and wetting).

We also have found it quite advantageous to use high temp (90% lead) solder balls (one in each corner) that do not collapse providing a column structure and near filleted solder joint as compared with the normal pumpkin shaped reflowed eutectic ball/paste joints. This is based on reliability studies done over some period.

There's much more, but as some folks say: "BGA'S aint no big friggin deal."

Earl Moon

The company whith which I am now under contract has more current and historical "hard" data supporting its decisions than any other place I've been before. Also,

reply »

Earl

you posted a message on 1-14-99 about converting to VOC free flux. Have you completed your evaluation? Could I get a copy of any data supporting your conversion to VOC free?

Thanks

WDR

reply »

| Earl | | you posted a message on 1-14-99 about converting to VOC free flux. Have you completed your evaluation? Could I get a copy of any data supporting your conversion to VOC free? | | Thanks | | WDR | | WDR,

We are awaiting the latest data translation and acceptance from the study. Our powers to be must approve the findings and allow publication.

Now that our company has become a no-name, 8 billion dollar IPO, we are going through some pain. It involves what to do with what and what we can say about it. Sounds like double talk to me, but I am hopeful.

I do promise to keep you all posted. However, I will say initial findings very positive for most all we do - though we cannot yet see ways to apply the stuff to everything. We are developing and transitioning two different solder pastes as well as spray fluxers and other complexities.

Please understand, we are doing what I call wierd stuff (stuff no one else in industry wants to do) and are very picky because of it. Plus, we are trying to keep it as simple as possible. You know, one or two paste types, one or two reflow profiles, minimizing selective wave soldering pallets, prolonging transition to OSP's, etc.

Thanks for the continued interest,

Earl Moon

reply »

| | On 1.27mm (0.050 inches) Standard Ball Grid Array's. | | The ball size diameter starts at approximately 0.025 inches. | | | | If the ball size decreases to 0.012 inches (due to the solder flowing down a via) is the solder connection reliable? | | | | This is a known Liquid Photo Imageable (LPI) Soldermask clearance violation. I have checked numerous seminar notes I have been to and a couple of hard bound technical books and came up empty as to any discussion pertaining to ball size and reliability. | | | | I am under the impression that if we have less than 50% of the ball diameter remaining that it would be a defect. | | | | Thanks | | | | | | | We allow no solder ball size reduction. We allow no solder paste/ball drain off into any hole type. We paste during production, of course, and rework. | | This may change when we have objectively determined certain micro via structures acceptable. This will only allow minimal solder ball size reduction. We consider reduction to be one of two things - either partial reflow (viewed with X-Ray as insufficient solder) or excess solder when balls are very close to bridging (a little more complicated but a function of too large pad sizes [design violations] and wetting). | | We also have found it quite advantageous to use high temp (90% lead) solder balls (one in each corner) that do not collapse providing a column structure and near filleted solder joint as compared with the normal pumpkin shaped reflowed eutectic ball/paste joints. This is based on reliability studies done over some period. | | There's much more, but as some folks say: "BGA'S aint no big friggin deal." | | Earl Moon | | The company whith which I am now under contract has more current and historical "hard" data supporting its decisions than any other place I've been before. Also, |

Earl, Thanks for the response. Andy

reply »