0 gap meshes/faces being fused by default
Hi, I have a presumably, zero tolerance model as part of my goal but the slicer software didn't allow me to have that. I wasn't sure where to ask, whether this is a bug or not either.
As the title stated, the slicer just somewhat combines the overlapping or zero gap faces by default even if i were to key in xy compensation in negative value. image below are what I meant, Left having a micron or less gap, while on the right they're either overlapping or zero gap on the model itself. Of course these models has been "split as parts" in prusaslicer. Files can be found here. Left is "New Test Print.stl", right is "Old Test Print.stl".
I actually tried in cura 4.13.1, the behavior of old test print.stl file is similar to the image below. I am somewhat wondering why the slicer behaves like this?
RE: 0 gap meshes/faces being fused by default
This should really be in the Prusa Slicer forum but.... 🙂
Theres a setting under Print Settings>Advanced>Slicing called Slice gap closing radius. Its a yellow dot setting so you need to be in Advanced or Expert mode to see it.
By default its set to 0.049mm. This causes issues with some of the tolerance tests too as they slice as a single area. Just set it to a much smaller number by adding a few zeroes after the decimal place like 0.000001. I'm not sure if this is one of those settings that zero disables.
RE: 0 gap meshes/faces being fused by default
This should really be in the Prusa Slicer forum but.... 🙂
Theres a setting under Print Settings>Advanced>Slicing called Slice gap closing radius. Its a yellow dot setting so you need to be in Advanced or Expert mode to see it.
By default its set to 0.049mm. This causes issues with some of the tolerance tests too as they slice as a single area. Just set it to a much smaller number by adding a few zeroes after the decimal place like 0.000001. I'm not sure if this is one of those settings that zero disables.
I actually tried the slice gap closing radius, even with the other option of making the resolution to 0 also resulting the same. I just tried with the cura v4.13.1 on the new test print.stl, the behavior is similar to the old test print.stl.
I don't think this is an issue only for prusaslicers, but on cura as well. maybe you should test out the models itself, don't have to print it.
On another note, I even tried to fix the object via nettfabb in prusaslicer, of the old test print.stl, no improvement either
RE: 0 gap meshes/faces being fused by default
I think you need to go pose this question at the github for the dev's to look at. https://github.com/prusa3d/PrusaSlicer/issues
With stl files theres probably a limiting number of decimal places for the math side of things when it comes to importing them into the slicer. The size of registers used in the code used will have a naturally limiting effect I think.
Also no hobby level fdm printer or even resin is going to have the printing tolerance for those sizes of gaps. Which is why no one is really worried about slicing such models that the printers cant reproduce anyway.
If you do have a need for an insanely close gap that cant be printed then the dev's are the correct people to talk to. Ask the question and if needed raise a feature request.
RE: 0 gap meshes/faces being fused by default
I think you need to go pose this question at the github for the dev's to look at. https://github.com/prusa3d/PrusaSlicer/issues
With stl files theres probably a limiting number of decimal places for the math side of things when it comes to importing them into the slicer. The size of registers used in the code used will have a naturally limiting effect I think.
Also no hobby level fdm printer or even resin is going to have the printing tolerance for those sizes of gaps. Which is why no one is really worried about slicing such models that the printers cant reproduce anyway.
If you do have a need for an insanely close gap that cant be printed then the dev's are the correct people to talk to. Ask the question and if needed raise a feature request.
Please excuse me for saying the following; I believe you're jumping to the conclusion too quickly thinking that i wanted to have a 0 gap tolerance printed parts. I modeled a 0 gap tolerance object for few number of reasons mainly; each printer has different tolerance, giving the user ability to customize their preferred tolerance and/or going more tighter tolerance than they ever been into. I stated this in the object link above in my first post here.
The intentions of 0 gap meshes/faces also is because some modeler doesn't state how many tolerance they've added into their parts, hence printing such object may turn out to be either too loose or too tight, of course, depends on the machine used as well. So my concept is to utilize the underrated xy compensation. And to repeat again, I did state xy compensation in my first post, just not being bold.
If you think I still should ask in github, I will. I figured I should ask here first before going there. Have a good day ahead 🙂
RE: 0 gap meshes/faces being fused by default
Its quite possible I misunderstood what you are trying to achieve. From what I understood you were trying to get it to slice the same with non joined geometry on both the model versions with no x/y compensation applied.
One of the things I was trying to get across is that your design software when generating the stl's may also be running into issues as there is usually a finite number of decimal places used when converting the native format used in CAD to that used for an stl file.
I looked at the stl files using a tool called 3d-tool which can do cross sectional display (very useful tool if you are on windows). The walls are rendered as interpenetrating along the curves. One thing I did note is that the overlap is in narrower sections on the New Test Print model.
Compound that in reverse when the slicer LOADS it in and even with resolution set to zero (so the slicer isnt simplifying things) then you are quite likely to see these issues when you are producing such models.
RE: 0 gap meshes/faces being fused by default
Its quite possible I misunderstood what you are trying to achieve. From what I understood you were trying to get it to slice the same with non joined geometry on both the model versions with no x/y compensation applied.
One of the things I was trying to get across is that your design software when generating the stl's may also be running into issues as there is usually a finite number of decimal places used when converting the native format used in CAD to that used for an stl file.
I looked at the stl files using a tool called 3d-tool which can do cross sectional display (very useful tool if you are on windows). The walls are rendered as interpenetrating along the curves. One thing I did note is that the overlap is in narrower sections on the New Test Print model.
Compound that in reverse when the slicer LOADS it in and even with resolution set to zero (so the slicer isnt simplifying things) then you are quite likely to see these issues when you are producing such models.
Actually I did xy compensation at -0.1 for the image in my first post above. I'll check the software you mentioned later. There's a very small gaps like 1 micron or lesser on the new test print.stl to circumvent the issues of the slicer just assuming any overlapping meshes are combined objects. However I believe the overlapping meshes has less than 1 microns or around nano meter range, yet keying in negative xy compensation does almost nothing to this upon slicing the object. I wasn't sure if this is intentional or what from what I see here is that the slicer will do the following in sequence on generating the gcodes:
- slicer will combine any parts that has overlapping meshes even less than 1 micron range
- generate gcode and coordinates for plastic extrusion
- xy compensation applied after no.2 above
What I initially thought or perhaps I kind of prefer was:
- generate gcode for individual parts/objects first
- xy compensation applied
- final output
However at the steps of what I preferred for the slicer's behavior, it would be very complex to implement and perhaps heavy on resource. Maybe I will open up an issue over prusaslicer github regarding this issue(?) and perhaps see where the direction goes later on.
You are asking the printer to work at close interference measurements. Whilst it's easy to work down to 0.5mm and reasonable to aim for 0.2mm clearances anything much finer has to be machined post-print. Worse, very fine curves have to be appoximated in steps and the paths generated will inevitably cross micron scale boundaries. In practice even a 0.1mm clearance is quite likely to print seized even though slicer is happy to work at that scale. As such the slicer may as well condense zero clearences into a single body.
If you want to design variable clearences or tolerances then it will be best to make them parametric with a reasonable default.
Cheerio,
RE: 0 gap meshes/faces being fused by default
You are asking the printer to work at close interference measurements. Whilst it's easy to work down to 0.5mm and reasonable to aim for 0.2mm clearances anything much finer has to be machined post-print. Worse, very fine curves have to be appoximated in steps and the paths generated will inevitably cross micron scale boundaries. In practice even a 0.1mm clearance is quite likely to print seized even though slicer is happy to work at that scale. As such the slicer may as well condense zero clearences into a single body.
If you want to design variable clearences or tolerances then it will be best to make them parametric with a reasonable default.
Cheerio,
Virtually speaking, yes, however we have benchmark objects such as tolerance coin to see your machine's capability. There are times that pre-assembled object file are better and can be simpler compared to an individual parts with assembly required, though each one of them has its own huge disadvantages.