How would you predictably control tin droplet behavior in the presence of swirling gas flows and plasma?