Tolerance - Dihedral and Tangency?

Mostly out of curiosity: Does anyone know what the dihedral angle
tolerance range is that will satisfy Pro/E's tangency requirements; e.g.
when is a surface intersection shown as tangent vs. sharp? Is it a hard
coded or configurable value?

TIA,
Jeff

 

Any dihedral angle of less than .5 degrees is considered tangent.

 

Many thanks.

 

: Mostly out of curiosity: Does anyone know what the dihedral angle
: tolerance range is that will satisfy Pro/E's tangency requirements; e.g.
: when is a surface intersection shown as tangent vs. sharp? Is it a hard
: coded or configurable value?
:

I came across this by accident recently, might be at least a partial answer to
your question. There's an option called tan_angle_for_disp, set by default to
..0216. Settable to anything, AFAIK. The description says its for display only and
is the angle at or below which surface patches show tangent lines. The rest of the
description says to enter an angle between 1.5 and 15 which seems to contradict
the default setting, unless this is just a way of slyly filtering tangent edge
display. Even their explanations create mysteries.

 

I came across this by accident recently,

Cool. Thanks, David. I'll have to fiddle with it when I get some spare
time.

I did play with a test object a bit. The results were interesting and
raised more questions (to which there are probably not simple answers).
The "test" was a rectangular block; section 20" x 5" extruded 2". Bisected
the top face with a datum crv (two point, on srf attrib, as were all that
follow) to create two 10 x 5 regions. Put a datum pt on one 2" vertical
corner edge. Created datum crvs from that point to divide two side faces.
Created a boundary blend using the three crvs and one edge (a simple ruled
surface; e.g. didn't specify G1 or G2 continuity) and cut the block with
it. By varying the corner edge datum point's offset from the top face (and
the resulting intersection dihedral angle range; 0 degrees on one end to X
degrees on the other end) it appears that there is some "averaging" going
on. Also some curious splitting / segmenting of the intersection curve
occurred in some ranges. The real explanations of what's going on are
probably addressed in someone's doctoral thesis.

Any way, I'll have to play with the variable some. It might be a good tool
for analyzing imported models; quicker than doing a dihedral analysis for
numerous intersections.

Often the case, idnit? 8~)