Holes-clearance for 0.4 nozzle
Question: does the Slicer distinguish between outside and inside shapes? To print a hole with exactly 5.00 mm I have to construct a hole with 5.10 mm to get the right result. I was expecting to design a hole with 5.45mm diameter (5mm plus .45 extrusion width @ 0.2 layer-height). Can someone explain me how the slicer calculates this?
I don't think there's any universal answer to this. I've found that holes depend on many factors, even such things as the width of the bead that's extruded. For any close-fitting parts I'll do a trial and error test print on such things as alignment pegs using the same filament and settings.
I would like to learn more about the dimensional standards and tolerances that are used for the machines and the slicers.
Holes almost always print smaller than designed. Different slicers calculate things differently so trying different ones is always a good idea. Some will do different things better than others, like supports may be better in one but another may do surface details better for example.
Most slicers will take extrusion width into account when planning the paths. After all you set that in your slicer settings. To keep things simple if you assume an extrusion width of 0.4 then the centre of that extrusion is going to be 0.2 in so for an external perimeter the nozzle needs to follow a line 0.2mm inside of the exterior shell of your model. At its basic level that's what it does. So your external dimensions are 'generally' correct. That is in an ideal world. Then you hit real world 🙂 In reality its not quite that simple as the extrudate isn't a perfectly circular 0.4mm, its squished down onto the layer below meaning it has a flat bottom and top and that the sides are further out. Theres some complex math involved to make things more accurate and I never did like math which is why I'm really glad the slicer sorts it out for me.
Part of it is that holes are not perfectly circular in the stl files. They are made up of short lines that give the appearance of a circle. When the model is sliced the print path will follow these lines so can often 'cut the corner' as it were on these straight paths. Then there is expansion and contraction of the material which is just the nature of printing hot plastic. Then there's other perimeters, infill etc and all this different plastic being laid down in all these different ways and things get complex. The more complex the geometry the more likely something will be slightly off. Also as slicer only work on a layer by layer basis (at the moment) they don't take into account layers above them where doing something slightly different might make a difference to the later layer.
There's also printer precision too. For functional parts that fit together you always have to design in some tolerance for both internal and external features where they mate. Some printers are better than others. That's why there are loads of tolerance gauge models you can print to figure this out.
My mk3 I have to use around 0.1mm tolerance but my cr10 is around 0.25mm . Nozzle size and even plastic type will/can make a difference to the tolerance required too.
Using a parametric capable design software where you can easily alter tolerance is a big plus when printing functional parts. Personally for holes I design them with sizes that give very close to what I want after using the print and see method, like your 5mm hole needing to be 5.1 and then for things like screws/bolts its very easy to ream them out to perfect size if I need to.
thanks for your thoughts. Indeed, there are many factors, I've found out that even different filament colours have different behaving, due to the pigment structure. But there must also be an algorithm in the slicer software to differentiate between inside and outside walls (see pictures, where the output is near the same diameter). I'm pretty sure slicer is calculating the outside dimensions and subtracting the extrusion width. PS: I've just analysed the g-code file and can confirm this fact.
CONCLUSION: So my "old thinking" of how walls are calculated (putting the nozzle exact to the dimensions coordinate) is wrong, instead, the slicer is calculating the position in respect to the extrusion width. Back to my "hole-problem": It's not necessary to deal with extrusion width's, just compensate minimal deviation for the "magic" influences other factors.