[Gulp35]

I was thinking last night about the shape of the impingement "dimples". When the bottom is cone shaped the water on the outside of the tube would hit the bp first and it would want to go toward the point of the cone. this would happen to all of the other minute columbs of water except for the one in the very middle which would, by this time, impinge on the columbs of water that have allready hit the cone.
this causes turbulence but turbulence above a thick boundry layer of all of the water.

Unlike the water impingeing on a flat surface which all of the water forces the boundry layer down increaseing heat transfer.

I am not a heat transfer specialist, but this sounds right, doesn't it?