Hi , I have issue no solder and hairline solder short during wave soldering my processor board. The issue happens randomly from board to board. I have changed my wave pallet opening and optimize my wave profile, during monitoring lot the result was good but when i release to mass production mode this issue happens again with yield lost around 0.4% from lot size 5k. Attached is the wave pallet opening and reject picture

 

That does not seem like a poor production yield, really. Years ago, among other duties I was the one who repaired the production soldering faults and developed the corrected processes.
But that job position did not last very long. My evaluation that it was operator errors was not appreciated at all, even though it was correct.

 

Flux. It's not getting good coverage. Possible the wrong flux? No, not at 0.4%. Four out of one hundred boards is not a bad run. It's 99.6% effective production.
Is the machine running a little too fast? Poor flux application can be the result.
I worked a rework line. The pick-n-place and wave solder machines were running too fast. We had 4 out of 10 boards that had to go for touchup or rework. Once they slowed the machine down a little they got much higher yields with far less rework.

 

What direction is the wave traveling relative to the board? You've got the PCBs that are protruding through the main PCB and I can see those disturbing the wave of they are parallel to it. You might try turning the board 90°, if possible. Then again, I have no experience in doing production work. so this is just speculation on my part.

 

Four out of one hundred boards is not a bad run. It's 99.6% effective production.

that math does not add up

 

that math does not add up

Oops! 96% is not a bad run. Good catch. I was thinking of 4 in 1,000. Actually 96% isn't all that great a run. That's 40 in 1,000, 400 in 10,000, etc.

 

At least in some organizations any flaws in the process lead to corrections aimed at preventing those flaws, often right away to avoid lost production.