LGA Solder Balls

I am having problems with LGA solder balls under the devices. Here are some of the details:

1. There are only 2 affected parts @4 locations.

2. The parts are LTM4616EV#PBF and LTM4608AEV#PBF.

3. We are using a ramp soak spike profile. 0-100secs=ramp to 150c. 100-200secs=soak@150c. 200-287secs=spike@220c.

4. This is a recent phenomenon. Approximately 3000 boards have been built with no issues.

5. Oven is calibrated.

6. No visible printing issues.

7. 5mil thick stencil.

8. We were underwiping stencil every other board, now we are wiping every board.

9. Placement machine does not appear to be over driving parts.

10. Solder balls are not apparent before reflow (x-ray wet board no issues).

11. Experiment Baked a lot of 50 boards before building and hte problem did not go away.

12. Experiment Baked all MS parts for 50 boards before building and the problem did not go away.

From this, my only conclusion is that it could be one of the following:

13. Moisture in the paste. (But why is it only affecting LGA devices?).

14. Paste is not thawed properly before use. (But why is it only affecting LGA devices?).

15. Aperature opening need review. (But why is this happening now after 3000 bds have been built?).

I need a sanity check on my logic. Any suggestions are welcome.

reply »