SolveSpace Logo SOLVESPACE -- parametric 2d/3d CAD

(you are viewing a thread; or go back to list of threads)

Self-intersecting mesh (by Eric Buijs)
I've created a simple inner thread in a cylinder. When creating an stl file I get the message that the mesh is self intersecting (see image). However when I check the same file in Meshlab or Cura everything seems fine and I can 3dprint it without problem. So is this a problem within Solvespace?

I'm using Solvespace 3.0 (Debian package 3.0rc2+repack1-3) on MX-Linux 21.
Wed May 18 2022, 05:46:58, download attachment Schermafdruk_2022-05-17_23-22-14.png
(no subject) (by Eric Buijs)
I've attached the .slvs file here
Wed May 18 2022, 05:47:56, download attachment inner_thread.slvs
move to 3.1 (by cmpxchg)
Hi Eric,

I can reproduce this screenshot -error with solvespace 3.0, but not with release candidate of 3.1-dfbefa60 anymore - no errors upon export, and prusaslicer taken both 3.0 and 3.1 generated STL files and slices them properly.
I can also remove your 'force nurbs to triangle mesh' checkbox on g005-lathe and exporting it still works fine.
I am running majaro arch linux and built solvespace-3.1 from source using the instructions.
Wed May 18 2022, 06:26:52
(no subject) (by Paul)

I see your triangle for cutting the helix extends past the surface it's cutting - good. It looks like that triangle has side length of 6, and then you constrain one turn of the helix to a distance of 6. This is causing a lot of "just touching" surfaces. You may want to slightly shrink that triangle. Also version 3.1 (RC) has an optional pitch parameter for helix groups - it's in the text window for the group - this will allow you to avoid the translate group entirely since you can set the pitch and constraint the helix length.

By changing the triangle height to 5.9 and setting my chord tolerance to 0.2 I got a few red lines without forcing to triangle mesh. The red lines are not on the surface and using "analyze show naked edges" actually makes them go away. Only problem is this changes the thread shape. You can tweak the triangle to keep the shape while preventing them from touching. I might work.

I've seen other cases where solvespace shows red and then specifically asking it to check for naked edges either fixes it or changes the number of bad edges detected. I'm not sure what's going on there.
Wed May 18 2022, 10:12:15
(no subject) (by Paul)
I've confirmed using 3.1 and the pitch parameter works great. No need to change the triangle dimensions at all. I deleted the final translate group, deleted the constraint that makes the helix one turn, checked the fixed pitch box, changed the pitch to 6, and then dragged the helix length (via the center construction line) and the spiral cut all the way down while dragging and never showed red. This works at CT of 0.2 percent and 0.1 percent setting, all without forcing to triangle mesh.
Wed May 18 2022, 11:25:47
(no subject) (by Eric Buijs)
First, thanks for the replies and great suggestions. I tried changing the triangle height (why didn't I think of this ;-)) and chord tolerance according to the advice from Paul. It reduced the number of problematic edges a bit but didn't fully eliminate them. Luckily I can produce a valid stl and 3dprint it without problems. The only thing that remains is that the slvs files are unusable for assembly in Solvespace which is a pity but not a disaster.

Also great to read here that v3.1 apparently solves this problem. Thanks for trying. Since my issue is not a showstopper I think I'll wait for v3.1 to come along in my distro. But I'm glad it's out there so I'm able to build it if I get in trouble down the line with this project or another.

Lastly I'm taking the opportunity to thank the developers for this awesome program and this community for all the help they provide.
Thu May 19 2022, 06:18:01
(no subject) (by Paul)

I've attached a version of this thread that should work in 3.0 even though I made it with 3.1. The key was to get rid of the translate group and your helix constraint, twist it manually 5 times, constrain point to plane for exaxtly 5 turns, then constrain the length to 30 (5 * 6). It should still open in 3.0 and if not you can repeat those steps in 3. The key to avoiding NURBS failure was to do all the threads with a single group ;-)

3.1 will probably be the most distro-friendly version yet. Ryan is actively tweaking things based on Debian packages with RC1.

You're welcome.

Also I'd like to add a link to your videos under "external resources" or similar on the main page here. Do you have a preferred URL?
Thu May 19 2022, 10:01:18, download attachment inner_thread2.slvs
(no subject) (by Eric Buijs)

thanks for the file. It opens in Solvespace 3.0 on my PC but it still has some problematic edges (according to Solvespace). I tried the method you described, manually create the helix by twisting a number of times. This is extremely CPU intensive on my, admittedly humble, PC making it hard to do. I believe that's the reason I came up with the translation when I started working with the helix. Well, at least it gives me another option when I run into serious problems in the future.

Meanwhile I'm excited about the upcoming v3.1 with the exploded view, the pitch parameter for the helix, the additional constraints and the ability to link stl files and much more.

Great that you want to add my videos to the Solvespace website. I think for the URL it's best to link to the Solvespace tutorial playlist (see below). In the coming months I'll try to update some of the older tutorials which were made with v2.3 and I'll upload some new ones.
Sat May 21 2022, 06:38:31
Post a reply to this comment:
Your Name:
Your Email:
(no HTML tags; use plain text, and hit Enter for a line break)
Attached file (if you want, 5 MB max):
© 2008-2022 SolveSpace contributors. Most recent update June 2 2022.