SOLVESPACE -- parametric 2d/3d CAD
Examples
Tutorials
Features
Download
Reference
Technology
Library
Forum
Contact
USER FORUM

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

a couple of comments (by R.Lucchesi)
Great program much better than trying to use other free CAD apps. The parametric make all the difference.

A couple of comments:
1) having an error message like "zero length segment" or "figure? not closed" is great. But showing some way how to fix it would be even better. You have that for many of the Constraint errors. But not for these more basic errors which can nonetheless be very problematic to remove.
2) When extruding a figure in a workplace and it cuts another extrusion, sometimes the cut leaves a closed figure and sometimes not. It's unclear what is wrong. I have tried to constrain the points of the new figure to the planes in the old figure but when working with assemblies of other parts and they come together where the new extrusion will go it seems like one of the parts can retain it's internal structure and the other is cut away.
3) Force NURBS surfaces to triangle mesh was able to solve #2 for me but I only found this by trial an error.

I must admit that I have only finished the 1st two tutorials and have come a long way. The 2nd one and constraining points on planes made all the difference for me, when working with assemblies and just working with extrusions.

I am having a problem and don't exactly know where it may lie. The dimensions I use in distance constraints are in MM and when I export the file as an STL and then print it on my 3D printer they are consitently off in 0.5mm or so. The printer manufacturer claims that they have a tolerance (at the printing level I am using) of 100 microns. So either the export process is inducing some distortion or my printer program is inducing some distortion or my printer is inducing some distortion. If it weren't so consistent I would think it was a printer problem. But it's very consistent. Fortunately, consistency can be worked around. But it may make my exports less usable by other printers. Not sure where this fault may lie and I have no way of validating the dimensions of the export files.

I have attached a Zipped folder with two files: test.svs and test.stl. Inside the "test.svs" file there is a extrusion with two holes vertically aligned that are 3.9mm in diameter and the distance between the two holes is constrained to be 27.5mm. I exported this file to "test.stl". If I were to print this file the holes would be closer to 27.0mm center to center.

Again the program is great, Kudos to you and your team and thanks.
Ray
Wed Oct 19 2016, 00:51:40, download attachment test.zip
(no subject) (by Jonathan Westhues)
(1) Maybe hovering/clicking the "zero-length" message should be equivalent to hovering/clicking the zero-length entity, for ease of selection? I'm not sure what would help with the "not closed" message, though.

(2,3) The NURBS Booleans aren't perfectly robust. The archives of this forum describe various ways to sketch that make such failures less likely. In any case, if you're exporting STL, then mesh Booleans work too.

When the NURBS Booleans fail, a message in the browser already notes that switching to mesh Booleans might fix that. I guess we could make that more prominent, bright red and bold or something.

I'm not sure what's wrong with your output scale. Is SolveSpace's export scale (in the configuration screen) 1:1? If not, then that's your problem; and if yes, then perhaps you can fix it there anyways.
Wed Oct 19 2016, 01:40:20
(no subject) (by whitequark)
(1) See https://github.com/solvespace/solvespace/issues/87.

Regarding the output scale, I've compared your file to one that I have created locally at 1:1 scale, and they are identical. Assuming your 3D printer uses open-loop positioning, I suggest checking the configuration of your machine axes, particularly a) whether there are any missed steps and b) whether the leadscrew pitch is set correctly.
Wed Oct 19 2016, 08:13:21
(no subject) (by R.Lucchesi)
Perhaps not closed would point to the two line segments that are the ones not closed.

As for zero length. What's the zero length entity?

So if you have checked the two files and they are the same that you created AND you can print them and they print according to the dimensions in the SVS file then it's a problem with my printer or printer program.

Thanks for all your help. I guess I will read more in the forum archives. Not sure where Mesh Booleans are indicated.

The export scale factor in the configuration file is 1:1.

Thanks for all your help.
R.
Wed Oct 19 2016, 21:40:29
(no subject) (by whitequark)
> As for zero length. What's the zero length entity?

A line segment where the endpoints are coincident.

> AND you can print them and they print according to the dimensions in the SVS file then it's a problem with my printer or printer program.

I didn't print them. I don't even have a 3D printer. I'm quite certain that the STL file is correct.

Mesh Booleans are implicitly activated if you click "Force NURBS surface to triangle mesh", so you already know that.
Thu Oct 20 2016, 02:44:32
So we printed the TEST.stl files on another printer (by R.Lucchesi)
i have a micro3d printer the other printer (i will need to find out but it's not a micro3d). In both cases although the .SLVS file says the holes should be 27.5mm apart when we print them they are 27mm apart.

There appears to be a problem in the STL export for SolveSpace.

You should be able to check the .SLVS & .STL files and print them on another printer if want to. But as far as I can tell there's a problem somewhere in the SolveSpace export.
Ray
Sat Oct 22 2016, 19:48:28
(no subject) (by Jonathan Westhues)
I'd suggest that you open the STL file in a viewer that you trust. I am nearly certain that you'll find the problem is either your 3d printers or whatever you're using to measure the parts: software bugs just don't usually work like that, and no one else has observed this problem in eight years.
Sat Oct 22 2016, 19:58:59
(no subject) (by whitequark)
Are you aware that thermoplastics exhibit significant shrinkage? Are you using PLA for your prints? Your discrepancy, which is 1.85%, is near the shrinkage coefficient of PLA, which is around 2%.
Sun Oct 23 2016, 00:28:11
We're using PLA (by R.Lucchesi)
I don't have access to an STL viewer that I trust.

As for PLA shrinkage, I will have to look into it.

I have not incorporated any shrinkage in our prints.
R
Sun Oct 23 2016, 02:03:49
We're using PLA on both printers (by R.Lucchesi)
I don't have access to an STL viewer that I can measure the inter-distance between the two holes

As for PLA shrinkage, I didn't realize that I had to account for that in my models. BUT it does make sense and it is consistent across multiple printers printing the same part with PLA.

That may explain what I am seeing in all my PLA parts.

Do you have a reference for the PLA shrinkage.

By the way, the holes are supposed to be 1.0607" apart which should make them be 26.94178mm apart and If I am using 27.5mm that's 1.020272% shrinkage. Which would be even closer to your 2% shrinkage on PLA.

Thanks for all your help in this. I think you nailed it.
R
Sun Oct 23 2016, 12:58:37
(no subject) (by EvilSpirit)
For measure STL, you can use http://notecad.xyz. Click ImportSTL and then you can measure point to point distances with Dist button. See attach for more info.
Thu Jul 5 2018, 00:21:59, download attachment import_stl.png
(no subject) (by Azial)
3D printed accuracy / tolerance depends on many factors, but not solvespace. The most impact has the slicer software you are using, followed by the material and then the physical printer properties.

E.g. when you want a cutout to fit a real object (e.g. a 3 mm screw in a hole in your 3D model), you have to make the hole bigger:
For loose fit, introduce a padding / radius of 0.3 mm,
For best fit, introduce a padding / radius of 0.2 mm,
For tight fit, introduce a padding / radius of 0.1 mm.

At least that is the case for a printer with a 0.4 mm nozzle using the dev version (> 3.0) of slic3r.
Thu Sep 6 2018, 03:35:24
Alexey, time to renew the http://notecad.xyz/ domain. (by ruevs)
It seems to have expired ;-)
Thu Oct 4 2018, 05:29:50
(no subject) (by EvilSpirit)
> It seems to have expired ;-)

I know. Will get back in 10 days, sorry for this ) Hard times.
Thu Oct 4 2018, 13:14:22
(no subject) (by EvilSpirit)
http://notecad.xyz works again and updated.
Sat Oct 13 2018, 02:57:39
Post a reply to this comment:
Your Name:
Your Email:
Subject:
(no HTML tags; use plain text, and hit Enter for a line break)
Attached file (if you want, 5 MB max):
© 2008-2018 SolveSpace contributors. Most recent update Nov 22 2018.