All you can really do with the model you originally wanted to print, is re-mesh or modify the mesh by trial and error until you get something that cura will spiralise as you want it to. Fusion 360 has some good tools for this. Try and aim for the triangles to be fairly uniform sizes and angles. Avoid long slivers and clumps of tiny triangles if you can.
37 minutes ago, mrender said:Trouble is the section isn't a continous curve, it is series of hundreds of straight lines without robust constraints on their endpoints because they are projections of hundreds of triangles at alsorts of angles, which themselves aren't particularly constrained to each other.
no, that's a closed NURB curve into the CAD software, what you see are the CV points...
There are a heck of a lot of them! Shouldn't need that many to define the shape. Have you got the native file? Which CAD package was it created on?
3 minutes ago, mrender said:There are a heck of a lot of them! Shouldn't need that many to define the shape. Have you got the native file? Which CAD package was it created on?
it is the result of an intersection, its normal the output is like that. Btw i uploaded the STEP file in a previous comment if you want to check it yourself
I see, would have to have a look at the native file to have any more thoughts on it.
26 minutes ago, mrender said:I see, would have to have a look at the native file to have any more thoughts on it.
do you have Rhino 7 ?
no sorry
1 hour ago, mrender said:Trouble is the section isn't a continous curve, it is series of hundreds of straight lines
No, that is not the trouble. Cura works with polygonal data, and because of that the cross sections will always consist of "hundred of straight lines". Cura inherently never sees curves. By your reasoning, no model could ever be spiralised in Cura.
1 hour ago, TonyCOD said:its 100% a closed loop
Did you see the part "without backtracking over a previously drawn line"? Are those lines that connect the different islands wide enough for two paths next to eachother?
Also, let's face it - and I am not being sarcastic - you are smarter than Cura. I am sure you can come up with models that theoretically should be spiralisable, but Cura may still fail to find the solution that is obvious to you.
Long story short: the "spiralise outer wall" function is not meant for these types of models. It is meant for solid models without multiple islands and without supports, and outer walls only. If you try to use it for something else, then yeah, the result may be inconsistent.
- 1
P. Slicer seems to be the answer after all.
- 1
6 minutes ago, mrender said:P. Slicer seems to be the answer after all.
do you have SW or Siemens NX?
also the vase mode stuff from Eclipson (they are gurus at this, so model errors can not be taken into account) its not sliceable by Cura, they provide g-codes and they are perfect
sent you a message
- 1
- 3 weeks later...
I tried another test, this works in others slicers, i do not know if the problems is my meshing procedure. Can somebody open this in a CAD, mesh it in STL and see if vase mode works??
Hi @TonyCOD
This is a stp file, not a stl.
However, I converted it to stl for you using Autocad fusion 360.
I'd expect a kind of vase model, but this is a Cube that cannot be spiralized -well unless you use "surface mode" single wall and without top and bottom..
Thanks
Torgeir
- 1
GregValiant 1,346
The model is close to being able to Spiralize.
You can see the white "layer start" indicator. It's there because on this layer an "island" formed that is disconnected from the main toolpath. Since Spiralize is on, Cura extrudes from one island to the other making the characteristic thick line between the islands. In addition, you can see that there are travel lines. If Spiralize had worked as you wanted there would be no need for travels.
I know this is hard to see but the diagonal lines running up to the top of the image are also connecting islands.
More starts and stops. There really aren't very many but every one results in a thick string (really an extrusion) running across air.
20 minutes ago, Torgeir said:Hi @TonyCOD
This is a stp file, not a stl.
However, I converted it to stl for you using Autocad fusion 360.
I'd expect a kind of vase model, but this is a Cube that cannot be spiralized -well unless you use "surface mode" single wall and without top and bottom..
Test_Step_AP203.stl 412.19 kB · 1 download
Thanks
Torgeir
Hi i know uploaded a STEP, if you read the thread and the comment you would have realized.
And you have the same problem as me, the mesh you created it's ill (see attachment).
Its not simply "a cube" if you watch it carefully every slice its a closed perimeter, in the same fashion of the other model i uploaded in the thread. In other slicers (i do not know if i can mention them) the vase mode on this model work, i want to see if the mistakes of Cura are due to mesh errors or something deeper. For now i state my last question again: can somebody mesh the step (which is clean of errors) in STL without errors?
2 minutes ago, GregValiant said:The model is close to being able to Spiralize.
You can see the white "layer start" indicator. It's there because on this layer an "island" formed that is disconnected from the main toolpath. Since Spiralize is on, Cura extrudes from one island to the other making the characteristic thick line between the islands. In addition, you can see that there are travel lines. If Spiralize had worked as you wanted there would be no need for travels.
I know this is hard to see but the diagonal lines running up to the top of the image are also connecting islands.
More starts and stops. There really aren't very many but every one results in a thick string (really an extrusion) running across air.
yeah, i see but in the model there are no islands, i checked it with a script very carefully, did you check the mesh before inputting it into Cura?
GregValiant 1,346
I repaired it using formware.co/onlinestlrepair and double checked with MS 3D Builder. It seems fine now.
There are two spots like this.
I noticed this as well. It looks like a layer is missing but I think the extrusion took a different path and is sitting inboard.
GregValiant 1,346
I see that. This is the repaired model in 4.13.1 and the original model that @Torgeir translated looks much like this with nearly the same missing features.
- 1
Hi @TonyCOD,
First, I'm sorry that I jumped in and disturbed in here -that was not my intention for sure.
So then to your problem, the advanced sliced cuboid, when we look outside this model, it seems to be nice, however, if we have a look inside (using a proper tool) - the model look horrible!
The inside "the cutted track's" is full of holes but the outlines is still there.. (-So it can be repaired sort of).
We may ask, why is it like this? I'll think you might know the answer.
I'll never have this kind of issue in the objects I've created and sliced, -never!
Cura show you all this, not having the auto mesh fill or mesh fix "fantastic" -that other slicers have...
If you have a detailed model, with lots of important issues inside, -what slicer would you use then?
Here's how your model (stl) looks from inside (using Fusion 360 full version):
This is just on part of one wall.
Here is a picture of Cura 5.40 B1 showing the same issue, not a coincidence.
I'll hope this give another impression of Cura.
Thanks
Torgeir
Edited by Torgeir
Corr text.
Hi @TonyCOD,
Here's the answer to your final question, your model without any issue.
Picture inside the stl file using Cura 4.13.1:
And here is the project file made with same version of Cura for an UM2E+
UM2E_Test_Step_Clean_h_0_15.3mf
Edit: Everything go wrong, added wrong project file this is the correct one.
This should be what you wanted, but took some time..
There is still some leaking joints..
Use 0.4 nozzle width and 0.15 mm layer height and all will be fine.
Well, -I did not use any mesh tools for this.
Also, in the "Mesh fix" deselect all the setting there.
My last input on this.
Thanks
Torgeir
Edited by TorgeirAdded text. Again.. Added wrong file.
- 1
It's okay to mention other slicers. You won't be censored or banned. 🙂
It's okay to say bad things about cura although some people might be offended and defend cura.
@torgeir this is a technique used by certain people who want full control over infill. Especially when making airplane wings where you want to set the ribbing (infill) in an exact pattern.
One of the problems is if there are too many points near the intersections. When cura slices a layer it has a list of unordered (unordered is crucial - it's a tragedy that STL triangles are unordered) line segments. It tries to make these into loops. Sometimes if lines are too close together it makes one loop into two. This is just one possible problem. I think somewhere you can set this value (the tolerance before it might make one loop into two at locations where the sides come very close together) but if not the solution is to make all the walls just slightly thicker - say 0.1mm.
Another problem you see with STLs is that a face has the wrong normal. Normals tell the triangle which side faces air and which side faces plastic.
The final common problem is holes - missing triangles in the mesh.
- 1
Recommended Posts
Top Posters In This Topic
25
10
10
8
Popular Days
Jun 6
18
Jun 5
10
Jun 21
8
Jun 4
6
Top Posters In This Topic
TonyCOD 25 posts
GregValiant 10 posts
mrender 10 posts
Torgeir 8 posts
Popular Days
Jun 6 2023
18 posts
Jun 5 2023
10 posts
Jun 21 2023
8 posts
Jun 4 2023
6 posts
Popular Posts
GregValiant
Hello. I thought I'd butt in as I have a different take on this. The 5.x versions of Cura are substantially different when it come to Spiralize. Those crossing strings you see are actual extrus
ahoeben
You don't have to "roll back". You can install and use 4.13 and 5.x side-by-side.
ahoeben
No, that is not the trouble. Cura works with polygonal data, and because of that the cross sections will always consist of "hundred of straight lines". Cura inherently never sees curves. By your reaso
Posted Images
mrender 26
Trouble is the section isn't a continous curve, it is series of hundreds of straight lines without robust constraints on their endpoints because they are projections of hundreds of triangles at alsorts of angles, which themselves aren't particularly constrained to each other.
Link to post
Share on other sites