Hi,
The wing was created as a loft with two profiles and two rails.
It also does not work as a reimported STEP.
It concerns 2 surfaces.
By the way, I have this problem more often.
Offset surface generates wrong result
I probably do not understand your question. All surfaces, including the ones I think you are indicating offset just fine.
John Hackney, Retired
Did you find this post helpful? Feel free to Like this post.
Did your question get successfully answered? Then click on the ACCEPT SOLUTION button.
There are even 3 offsets wrong.
The bottom of the slots are not all the way parallel to the outer surface.
Accepted solution
Ok, you are talking about the Offset Faces, now I understand. I see the issue now and I am not arguing that it is an issue. You can get around it using an Offset Surface and Thicken Cut the area. I only did one, to demonstrate. Model attached.
John Hackney, Retired
Did you find this post helpful? Feel free to Like this post.
Did your question get successfully answered? Then click on the ACCEPT SOLUTION button.
I suspect that this is a faceting problem, not a geometry problem. That is, the geometry is OK, but the resulting graphics is incorrect. I have not proven this theory, though, but I will try to look at it later. Does not make it less of an issue, I realize, but I suspect/believe that the geometry is actually fine...
Jeff Strater
Engineering Director
@jeff_strater schrieb:Does not make it less of an issue, I realize, but I suspect/believe that the geometry is actually fine...
Just look at the STL
The STL is generated from the graphics data. So, if the graphics are wrong, so will be the STL.
Jeff Strater
Engineering Director
@jeff_strater schrieb:The STL is generated from the graphics data. So, if the graphics are wrong, so will be the STL.
Whose graphic is wrong in this case. My graphic or Fusion's?
So far, we have always been able to demonstrate that there is a graphics error, especially with an exported STL. Only in this case it should be an exception?
Have you been able to check the part in the meantime? All I need is a usable part that can be manufactured.
But if the drawing also clearly shows the deformation, it is clear to me that it is not a graphics problem.
Interesting, I trimmed the ends using off-set place from XZ Plane, ~0.4mm of each end and it fixed some of the defects but not all
@jeff_strater I think there is more to it than "meets the eye". Pun intended 😉
We can change the graphics or tesselation density of an object in the viewport by changing the Display Detail Control from "adaptive" to fixed/high. I have not tested this, but I doubt that changes anything when said object is exported to a .stl file
We can also change the .stl output parameters when exporting the .stl, but I doubt that will change the detail level in the viewport display.
I would think that your statement only applies to default settings, so the .stl output options in their default and the display detail control is set to adaptive.
This has come up a number of times over the years and it would be nice to know how this really works.
I'll try to address a few topics in one post. But, first, the early indications are that my theory that it is the graphics (actually the faceting/tessellation) is not correct, and that @wersy is correct, and there is an actual geometry problem here. More as we find out more about this model.
"Whose graphic is wrong in this case. My graphic or Fusion's?"
If it had been a graphics issue, it would have been Fusion's. And, even, "graphics" is not really accurate. It is really "faceting", but that term is not widely understood, so I usually just say "graphics". Faceting is the process where a 3D model is converted into triangular facets, usually for display in a graphics window. Graphics cards, today, only understand triangles, so this is necessary to be able to see the model. But, this faceting is always an approximation to the real underlying surfaces. My theory that this was faceting is because I have seen many models that show these kinds of issues that did turn out to be just a faceting issue.
"What is the connection between graphics and STL?" (this was not directly asked, but implied by "I think there is more to it than "meets the eye"", and "it would be nice to know how this really works").
STL is also a triangular mesh, so the same exact code is used to convert the model to graphics and to STL. By default, I think, the same default level of approximation is used for graphics and STL faceting. Both are possible to change: With the set level of detail on a body (for graphics), and the accuracy settings in "Save as Mesh". Sometimes (not always), faceting errors can be corrected by varying the parameters for the faceting.
hope this helps.
Jeff Strater
Engineering Director
@jeff_strater wrote:
..
If it had been a graphics issue, it would have been Fusion's. .
I don't think it's a graphics issue. I don't know how often I have repeated this on this forum but here goes again.
When working with curves with variable radius, check your curvature! Always!
I am convinced this would not happen if the curvature would be smooth in that area.
If you try to use the press-pull tool in that area only, it will return an error message.
@TrippyLighting schrieb:
I am convinced this would not happen if the curvature would be smooth in that area.
I wish that were the solution.
On the contrary, it is getting worse. Now all the slots deform the wing.
Previously I used a fit point spline from a .dat file.
This time I used control point splines with minimum number of points.
Accepted solution
My apologies, I spoke too soon!
I exported the geometry in STEP format and opened it in SolidWorks 2023 and the model shows the same artifacts.
I wonder if this is a limitation of the solid modeling press/pull tool. If you create the same geometry using surface modeling, this work just fine:
Just a quick status update on this issue: No real answers yet, other than to validate that there is a geometry issue caused by the Offset Faces tool here, and we are continuing to investigate it. For a bit deeper bit of info on the result (not the cause): What is happening in the failed areas is: the modeler is collapsing what should be two vertices into one "tolerant vertex" that has a huge tolerance (3 mm). In the image below, the red circle shows this collapse, while the green circle shows the correct results. Very strange...
Jeff Strater
Engineering Director
Accepted solution
Thank you @jeff_strater for the comprehensive explanation.
And thanks @TrippyLighting for your effort. With this method, you also get a perfect result.
I was more familiar with @jhackney1972 method, so I used it.
Now I just thought of a method that is even simpler than "offset plane".
It is "extrude to object".
Now I may edit my wing one more time....
Fortunately I only worked on one third of the wing 🙂
I really wonder why I didn't think of it much earlier, after all I have used "offset to plane" many times before.
In any case, I was able to learn a lot again.
Many thanks to all of you!
