Vase mode issues again. Glitches from nothing

[nerys]

OK so basically I have micro defects in the model that the modelling program itself can't even see..... (those defects do not exist in tinkercad)

To slice this with PS you need to alter PS a little. the biggest change you must make (and is probably why blender is throwing a fit maybe as well??) those cuts are 0.01mm thick. in prusa slicer you have to goto print settings and then advanced. you need to change min radius to 0 or it will delete all or most of the "ribs" (fins) in the model and break vase mode hard core (I have sliced some works of art in PS before I found that settings)

THEN at the top. the 1st 3rd and 4th options must be set to 0.04mm manually and layer height must be fixed at 0.2mm and mix maxed at 0.2mm (basically you have to "disable" all the adaptive stuff for layers and extrusion or it throughs a fit around the 0.01mm cuts.

Once you do that I get a nice clean slice in PS (which is odd because usually S3D gives me the clean cuts and I have to fight with PS don't even THINK of trying cura. wow does it butcher them hard core. I have yet to figure out how to make it slice them.

The threaded sections are "known good" parts that I can slice fine in S3D and PS. I just copy and paste those bits into each new design I come up with.

the "new" bit here is the pumpkin shell.

Here is the PS project file (not sure if this is equiv to the factory file) I did find one glitch at layer 230/231 there is a missing segment that I had not noticed before.

[nerys]

THAT defect does not exist in the model. it is being "created" by S3D when it slices the model. to see what I mean try ROTATING the model. all the glitches are suddenly in different spots. that place where it creates an island as you should in your screen shot? that point is 0.8mm away from the outer surface of the model and I do not see that defect in the model when I make it transparent and zoom into those spots.

IE it does not exist. S3D seems to be "making" them when it generates the slice. no doubt something in the model is CAUSING S3D to make them but "I" can't find it. I don't even think tinkercad knows what normals and flipped normals are..... How do I address them?

YES this is vase mode compatible. JUST LOOK at where the red travel moves and retractions are. there is nothing THEIR for them to exist. just trust me its vase mode compatible.

Still you might want to check the model file again. Some parts aren't properly seperated from the
outer shell, causing additional loops.
sepvase.PNG


In other places there's too much seperation, causing closed perimeter loops:
sepvase2.PNG


Also this model reports as 2 shells...

[nerys]

First some amazing analysys. thank you. not sure how I can fix it though :-) hehe

NOW tinkercad does have this issue. if a model gets too complex it starts to FABRICATE defects literally out of thin air (literally taking chunks out of you model or adding chunks that literally don't exist)

the only way to fix that is to start a new work space with less in it or simplify the model. (use more primary parts etc..)

For example to make the ribs for assymetrical models (like the pumpkin) I have to create a MOLD of the entire pumkpin itself. Then I have to use that mold 15 times to "cut" 15 of the 16 ribs to be conformal with the skin of the pumpkin.

This means my model would contain that pumpkin 16 times. 15 of the ribs will have it as part of their makeup and the one visible as well (the 16th rib is a through and through so it does not need one)

This is not possible in TK. not easily. it would throw a fit at such complexity. so I used PS to "simplify" the pumpkin from its original 141mb file to a 2mb file. THEN once I finished all the ribs for the model I export it as an STL and then "reimport" the STL back into TK.

NOW its just a "simple object" instead of 15 ribs with a full pumpkin as part of each of those ribs structure)

I wonder if this model is just too close to the complexity limit. "I" do see some "micro" lesions on on the model. we are talking smaller than the extrusion width defects that the slicer SHOULD ignore since they are below resolution threshold. but maybe that is how these "flipped normal" manifest in tinkercad ???

Here is a picture of a coupe

THis was annoyingly hard to capture (the second one) as TK does not like it when you zoom into sub millimeter distances. the 3D nav UI/Controls do not like it. :-) whatever that is its less than 0.01mm in scale.

There's actually some intersections in the model... don't know why NetFab et al don't get triggered properly by
this, but this would explain why I can't get this sliced in vase mode in any slicer...
.....
....
Would love to see how this was sliced in vase mode in PS...

[parallyze]

OK so basically I have micro defects in the model that the modelling program itself can't even see..... (those defects do not exist in tinkercad)

To slice this with PS you need to alter PS a little. the biggest change you must make (and is probably why blender is throwing a fit maybe as well??) those cuts are 0.01mm thick. in prusa slicer you have to goto print settings and then advanced. you need to change min radius to 0 or it will delete all or most of the "ribs" (fins) in the model and break vase mode hard core (I have sliced some works of art in PS before I found that settings)

Thanks, I already had the slicing gap radius included - that's not what's causing the problem for me.... it is..

THEN at the top. the 1st 3rd and 4th options must be set to 0.04mm manually and layer height must be fixed at 0.2mm and mix maxed at 0.2mm (basically you have to "disable" all the adaptive stuff for layers and extrusion or it throughs a fit around the 0.01mm cuts.

...the extrusion width. That shouldn't really affect slicing a solid object in vase mode, but it does:

...using 0.4mm:

...and using 0.6mm:

Which leads me to the question:
Did you model this as a solid object or is it intended to be designed using thin walls, like certain
airplane models which pop up all the times in forums because they require specific parameters because of their... uhm... special
design?

From the looks of the STL file I'd guess it was modeled using solids and the cuts are simple 0.01mm cuts, like I personally
use them for vase mode prints, too and as mentioned in the example above.

The threaded sections are "known good" parts that I can slice fine in S3D and PS. I just copy and paste those bits into each new design I come up with.

So it's only the center "pumpkin" part? Maybe you can fix those intersections by loading the object into something
like MeshMixer and use the smoothing brush across the surface on very low intensity levels, so it doesn't alter the
geometry too much.

Or maybe reducing the mesh resolution might work, not sure if this might destroy the tiny cuts...

[nerys]

I downloaded the airframe without the threads and it only has a couple of error spots now....... WTAF how does something 100mm not even printed yet generate so many errors no where near it? how does that even compute?

Part of the reason PS sometimes does a better job is it will alter the model to "make" it work. for example I have a 3 angle intersection in a model. this gets dicey when one of those "faces" is creating layer lines parrallel to the build surface since it can create orphans that get "pinched" off the model below its print threshold creating a wee tiny little island.

PS will simple DELETE the orphans all together. Restoring vase pathing. S3D does not. it tries to print that wee little .1 x .1 mm orphan that it created because it decided to pinch it off as being too thin :-) hehehe

THEN. when it does see a problem WHY does it LEAP around the entire damned model to start again somewhere else instead of just moving the 0.01mm past the defect and KEEP GOING ??? these huge movements and "restarts" to the vase algo is what are creating the problems.

For example looking at this factory file layer 210 WHY in gods name does it STOP at the seam. move 720' around the entire damned model then violate my do not pass open space rule (maybe that path is over 999mm??) just after is passes one of those "notch outs" that do not exist in the model but maybe that is a flipped normal and this is how S3D see's it ?? and then to just go back where the hell it was 0.01mm away. this gratutious traveling "IS" what is breaking the model in print. if it would just JUMP the 0.01mm and be done with it..... I could just ignore the defects.

But when it recalculated the vase algo rise rate anew at this break SOMETIMES it does a squish (too low) which gives a solid model but an annoying but acceptable little bulge in the outline but SOMETIMES it goes the other way (too high) like in the PS glitch. and leaves half a layer unprinted. that is not acceptable.

ANother oddity. WHY did it jump here? the slice at this spot looks clean. WHY did it STOP at that corner on the right jump the 0.01mm gap in a travel movement (which its not supposed to do as I told it not to do that) and then run around the entire damned model for ...... reasons ? "WHAT" is in this spot that it felt it needed to stop?

Saw your new post added comments on that here

I do solid and then "remove" cuts to link all the faces.

I don't think I can do that with meshmixer. I think these artifacts are being added by tinkercad so once I brought it back into TK to add my parts I think it would add those defects (randomly) again. I can try it though can't hurt except to use sometime.

and yet most of the errors went away without the threads even though the threads are no where near the errors.... I just don't get that. why is S3D "adding" to the model features that are not there? is this a normals thing?

Maybe the models are not the same at all. maybe when I export the pumpkin airframe without the threads its literally NOT the same model (the common parts) but when I add the threads I am increasing the complexity and this effects the entire mesh it generates when it saves to STL......... interesting. a bit more export at last known good and re-import simple models might reduce the complexity enough to stop the errors??? I will have to play with that.

I just wish S3D was smart enough to go. hey. maybe just maybe this 0.00001mm defect is just a glitch. let me ignore it and just KEEP MOVING. of course as someone who barely understands what "code" means its easy for me to say that :-(

[parallyze]

NOW tinkercad does have this issue. if a model gets too complex it starts to FABRICATE defects literally out of thin air (literally taking chunks out of you model or adding chunks that literally don't exist)

the only way to fix that is to start a new work space with less in it or simplify the model. (use more primary parts etc..)

For example to make the ribs for assymetrical models (like the pumpkin) I have to create a MOLD of the entire pumkpin itself. Then I have to use that mold 15 times to "cut" 15 of the 16 ribs to be conformal with the skin of the pumpkin.

This means my model would contain that pumpkin 16 times. 15 of the ribs will have it as part of their makeup and the one visible as well (the 16th rib is a through and through so it does not need one)

Whooha... that does sound ... incredibly comfortable to work with. ;-D

I'm not using TK myself very often, but running multiple boolean operators on objects with high polygon counts is
something I've seen fail everywhere, no matter if Blender or 3ds max or meshlab... the moment you import high
poly geometry inside CAD there's always a chance for things like this :-/

I wonder if this model is just too close to the complexity limit. "I" do see some "micro" lesions on on the model. we are talking smaller than the extrusion width defects that the slicer SHOULD ignore since they are below resolution threshold. but maybe that is how these "flipped normal" manifest in tinkercad ???

We can't really tell S3D anything about the resolution of the resulting tool paths and what to ignore. What you can do
is a really wiggly workaround. Increase the cut thickness to 0.05mm and in S3D use the horizontal compensation using
+0.02. In theory this should leave a 0.01mm gap at the important locations and maybe ignore/bulge out the tiny tiny
intersections.

I tried to use the reduce modifier in mesh mixer, this seems to fix the outside a bit but in return starts eating up the
internal cuts in some places. Didn't test much, just a reduction of 20%, so results may vary a lot. And it's probably best
to start without the threads inside the STL.

Using the Mesh Reduction in S3D doesn't seem as effective as the reduction in MM here.

[nerys]

Hmmmm can you explain or link me to something that explains this HZ Comp function and exactly how it works?

Can I alter "gap" without altering "PATH" integrity and without changing the overall dimensions of the model? (these parts have to interface other parts)

For example on ONE printer I can't print these models. the SEAM does not fuse. somehow someway that printer prints that 0.01mm gap and RETAINS the gap. (same gcode) IE with little effort I can pull the seam right apart while when I use my other printers I can DESTROY THE PART and that seam does not fail. because its passing 0.01mm away from another extrusion it "fuses"

Can I tell S3D to cause a collision and IGNORE IT completely? IE I want that seam to overlap but NOT break the vase tool path. would HZ Comp do this or would that recalculate the tool path? IE I really need that seam to FUSE solid this is a rocket and that airframe is pressurized for a moment when the ejection charge fires. if that seam is not fused it blows open at the seam.

I know slicers can do this because I can do it in Ideamaker. but its an accident. ie not intentional. I don't know how to command it to do it. on one particular model the SEAM instead of being 2 walls 0.01mm apart they OVERLAPPED right side went left left side when right like overlaying pages of two phone books interleaving them. that would be an insidiously strong seam !!! hell it would not be a seam anymore but still slice like one

If I could do that on purpose IE tell the slicer to do that. WOW I could probably make the wall gap dimension HALF what I am using now and still be strong enough (fast lighter less plastic less print time etc..)

[parallyze]

I don't think I can do that with meshmixer. I think these artifacts are being added by tinkercad so once I brought it back into TK to add my parts I think it would add those defects (randomly) again. I can try it though can't hurt except to use sometime.

Hmm... if those defects come from TK it might be worth testing another program. But that's hard to tell, somebody who has never
used something like Blender before will probably run away, screaming wildly... I've never used F360 on imported geometry, but that's
at least another variant which would be free for personal use/testing.

and yet most of the errors went away without the threads even though the threads are no where near the errors.... I just don't get that. why is S3D "adding" to the model features that are not there? is this a normals thing?

Try not to get stuck into that rabbit hole... many programs will behave very differently and often there is no real solution/answer to questions available. File formats like FBX which will irritate almost anything but Autodesk products or DXF files from Inkscape which
will work great in FreeCAD but not in Inventor. There's many tiny little details which can go wrong. And us using a file format from the 1980s (STL) isn't really helping here, the segmentation taking place might also cause tiny imperfections when approximating all the curves on an object... ^^

Maybe the models are not the same at all. maybe when I export the pumpkin airframe without the threads its literally NOT the same model (the common parts) but when I add the threads I am increasing the complexity and this effects the entire mesh it generates when it saves to STL......... interesting. a bit more export at last known good and re-import simple models might reduce the complexity enough to stop the errors??? I will have to play with that.

I'd try to import the pumpkin in Meshmixer and reduce the resolution before doing any cuts. Maybe TK is already reducing the resolution, don't know if the source file had a much higher count that the 32k polys from the STL. But it's definitely worth a try, I'd say.

I just wish S3D was smart enough to go. hey. maybe just maybe this 0.00001mm defect is just a glitch. let me ignore it and just KEEP MOVING. of course as someone who barely understands what "code" means its easy for me to say that

Can only guess here. But I do think sometimes slicers simply only recognize this:

And from a slicers perspective you're not really running into trouble until reaching the top .... ;-D

Edit:
Just rememberd - MeshMixer does also include boolean operators, so maybe you can
design your parts in TK all on their own and do the boolean operations in MM afterwards?

[parallyze]

Now, my head hurts... lol

You're welcome... xD

Try to focus on the outlines. Once you see the donut transformed to a reverse "C" with both ends
almost touching things start to get easier

[parallyze]

Hmmmm can you explain or link me to something that explains this HZ Comp function and exactly how it works?

I've written something about horizontal size compensation not too long ago, I think... here it is:
viewtopic.php?p=57138#p57138
There's another post a bit later in that thread showing this on more complex parts and the difference
to scaling x/y.

Can I alter "gap" without altering "PATH" integrity and without changing the overall dimensions of the model? (these parts have to interface other parts)

Hmm... I guess the seperation distance will always be calulated based on real gap width (0.01mm) + half of the actual extrusion width, if perimeters overlap due to horizontal size compensation they will likely "fuse" in the slicing process, killing vase mode compatibility.

For example on ONE printer I can't print these models. the SEAM does not fuse. somehow someway that printer prints that 0.01mm gap and RETAINS the gap. (same gcode) IE with little effort I can pull the seam right apart while when I use my other printers I can DESTROY THE PART and that seam does not fail. because its passing 0.01mm away from another extrusion it "fuses"

Hmm... any chance that printer is running a firmware supporting linear/pressure advance and the corners might be a bit thin, giving a nice angle to seperate the walls more easily? Or maybe there's other differences and/or slight underextrusion going on? Are the thread fittings any different from the other printers?
I don't know what the lowest distance is S3D will still reliably detect, maybe going as low as 0.001mm might work?

Can I tell S3D to cause a collision and IGNORE IT completely? IE I want that seam to overlap but NOT break the vase tool path. would HZ Comp do this or would that recalculate the tool path? IE I really need that seam to FUSE solid this is a rocket and that airframe is pressurized for a moment when the ejection charge fires. if that seam is not fused it blows open at the seam.

I've never really looked if perimeter overlap on the advanced tab will also have an effect on the outside perimeters touching. I guess it's only used for the internal overlapping parts of walls, not the outsides.
A mix of all the features might work.. but that's really hard to tell from a distance, some ways just take a lot of trial and error. A combination of increasing the cut width and using the horizontal size compensation in addition to slightly overextruding might also work, but that'd require a lot of scaling and it's likely better to get the pumpkin part to slice without problems first, so you can focus on the seam.

I know slicers can do this because I can do it in Ideamaker. but its an accident. ie not intentional. I don't know how to command it to do it. on one particular model the SEAM instead of being 2 walls 0.01mm apart they OVERLAPPED right side went left left side when right like overlaying pages of two phone books interleaving them. that would be an insidiously strong seam !!! hell it would not be a seam anymore but still slice like one

Overlapping every other layer in vase mode would indeed be a nice feature. Any chance IdeaMaker is using some kind of "slicing resolution" and depending on the position of the object this "bug/feature" gets triggered? I've seen something like this happening in Cura, sometimes moving an object by +/-0.01mm would make some cuts disappear. But only some, not all.