Non smooth surface, with rebuild surface and curve?

i did it with ,
extract surface and split non smooth part of surface and the edge too, then
rebuild the edge part change decrease high point count which make the surface no smooth, but the curve is slightly change with 0.0006 deviation ( my absolute tolerance is 0.001)
after sweep2 and join all surface back, i got naked edge at the part I try to solve the non smooth surface part,

can I ask, how is the right way to solve the non smooth surface? (because in mesh polygon has crack with it)
I confuse with rebuild function , it have change the curve slightly that make naked edge problem, how should I solve this non smooth surface please , thanks :slight_smile:

Can you post the .3dm file? You can use the vertical arrow icon above where you type your message.

first surface non smooth

2nd surface non smooth

3dm file
no smooth surface.3dm (1.1 MB)

The green surface is folded back on itself. Isocurves cross, and with control points turned on the dotted lines connecting the control points are tangled.

How was the green surface originally created?

while i try to array the control point of the surface 1, the surface 2 is seems better, but the bottom of surface 1 still cant do, how can I delete the control point ?with what command ? can I delete the bottom of surface 1’s control point ?

with fillet surface, use the profile of the fillet

and how can I select the nextcontrol point ?

Run Rebuild edges on the main body trimmed surface. That fixes the main issue. Then rebuild the radius surface sweeps as degree 3-4 points. Then match the surfaces etc. Degree 2 surfaces are all kinds of trouble.

Your orange surface is a bad object. Rhino should identify this surface as a bad object. Despite years of requesting that this sort of bad object be flagged McNeel prefers to leave these booby traps for users to stumble over. Rhino version 2 would identify this as a bad object but McNeel has refused to remove this serious bug from all subsequent versions.

If you run offsetsrf on the orange surface you will get the same sort of self-intersecting result. Bad objects tend to lead to creating more bad objects.

The reason the orange surface is bad is that it has internal discontinuities due the fully multiple knot that is located at the point where your fillet goes bad.
One way to fix the orange surface is to remove 2 knots in the V direction at that knot line. That will make the surface internally smooth without changing the shape very much.

There wasn’t a whole lot of user input at their request here:

I am interested in hearing any input on how and when you would want to perform the tests.

It is not the responsibility of users to validate geometry.

One of the reasons Rhino has a reputation for being such a minefield is that developers have continually resisted helping users identify where the mines are so that users only find the mines when they blow up and wreck their models.

Internal discontinuities in curves and surfaces is a well known bad geometry that causes Rhino commands to fail and/or create more bad geometry. It would be trivial for Rhino to identify these objects and notify the user.

The test for bad objects should happen when the bad geometry is created. However any reporting of bad objects should wait until just before control is returned to the commandline. A command like import should wait until all objects are imported before creating a pop-up window that notifies the user that bad object(s) were created.

hahah :slight_smile: thx stratosfear , after rebuild edge, while using match srf it say cant be run, because my surface is trimmed even already shrinktrimmedsrf :slight_smile:so i use sweep2 :slight_smile:

hmmm… i am just not expert user … may be beginner :slight_smile: i think as long there is a trick to know , there is good for now, there forum exist to give good info to use the product to surpass this bug condition, may be that bug is good bacteria to let us know our body has problem :stuck_out_tongue: to fix, or may be there is some calculating has to let it be there , which has been managed in combination with other complicated command or process :slight_smile:
as long there is a way, some problem is good to let us know that is now perfect life :stuck_out_tongue: