"Jorge R. Frank" wrote in message
...
Normally, the boundary layer trips
to turbulent well after the period of peak heating but rough surfaces can
result in early transitions. A transition prior to Mach 21 can cause
vehicle damage and a transition prior to Mach 24 can cause loss of
vehicle. The improvised Columbia repair would likely have gone turbulent
right from the beginning of entry (Mach 25), exposing the RCC panel and
the trailing black tiles to the superheated air. Columbia's damage
occurred at just about the worst possible location since the shock from
the nose cap intersects the shock from the leading edge between RCC
panels 5 and 13, depending on Mach number.

That's what some of these boneheads don't understand- you can't just use a
spatula and smooth it by eye. It has to be smooth to an extremely fine
degree to hold off (not prevent) boundary layer separation.

I'm afraid it's not something that Real Men (tm) are going to be able to do
off the cuff hanging on the end of a robot arm.