On 2010-07-28, DoN. Nichols wrote:
On 2010-07-28, Ignoramus8984 wrote:
Wrote another G code generator function.

Given x and y of the cone center axis, z of the bottom and top, top
and bottom radius, and a few other parameters like finish quality, I
can mill out a vertical cone. I first rough it out quickly, and then
mill the inner surface to desired quality.

How about modifying that to mill a projecting cone to fit it?
As long as you have the cone generation capability -- just add a switch
option to define whether it is male or female. (In other words --
generalize the program as much as possible.)

That would be the next program. But it is not really the same.

Ideally I would like to find a way to mill any pocket given by a
reasonably friendly predicate function p( x, y, z );

And then test the two against each other -- even in wax.

Hmm ... what happens if you make it larger at the bottom? It
would need a ball end mill which is more of a ball -- not just the
bottom half of the mill. And you would need to be able to tell it what
the diameter of the shank above the ball was so it could predict
interference.

I do not really need that

Of course, with such a mill -- you would have a lot of
deflection with a larger diameter cutting section and a skinny shank.
They probably don't make those for metals anyway. :-)

Oh yes -- do you also have the ability to specify conventional
or climb milling? If so -- do the heavy work in conventional, and the
finish cuts in climb which makes a cleaner surface (especially in
aluminum. :-)

sure

i