I’m not sure I understand the scope of your paper. Is it that you are designing for machines with a tolerance of zero? Has such a machine been invented yet? Or, is the assumption that a slicer slices to CAD dimensions? Because, it most certainly does not. The slicer slices for a particular model of machine and other parameters.
If one designs a hole depth of 2.05 mm, and uses, for example, the standard Bambu Studio profiles, the slicer (Bambu Studio) will lay down a layer with Z-position of 2.00 mm and the next layer at 2.20. The slicer will have to choose one or the other, resulting in a hole depth of 2.00 mm or 2.2 mm, certainly not 2.05 mm.
That is not to say that one cannot nail a particular feature (a hole depth of 2.05 mm), but other parameters will have to change. For example, one technique is to introduce 3 previous layers at a layer height of 0.15 mm. This will create an absolute Z-position offset so that the top of the critical layer falls exactly at the 2.05 mm hole depth.
This is CAD to the tolerances of the manufacturing process, which is always the case. It is not inaccuracy. I too have to design my models to the tolerances of the 3D printers on which I will be printing. And yet, using a standard Bambu Studio profile (0.20 mm layer height), I could achieve a 20 mm test cube, measured at 20.00 (x) x 20.00 (y) x 19.98 (z) mm. That’s pretty darn accurate in my little CAD/CAM world. It’s probably showing that I could fine-tune my 1st layer; and, all that within the tolerances and accuracy of my callipers.
I’m sure I_Built_A_Thing will respond himself, but I’ll point you in the right direction:
The points you’re raising, as well as the question you asked, are covered in the first minutes of the video. Check out the first post. If you can’t watch the video, you’ll find a summary in text form in the first post as well, which answers many of your questions.
Of course, you’re right: Every manufacturing process introduces some degree of inaccuracy, but we’re talking about the slicer here, not FDM itself. FDM inaccuracy would be a separate topic and worth its own discussion, but it isn’t directly part of this topic.
As for the rest, I_Built_A_Thing will surely answer them himself.
Sharky, I’m not speaking of accuracy. I’m speaking of precision. Any slicer slices to the precision of the underlying printing environment. In the Z-direction, the common layer height for a 0.4 mm nozzle is 0.2 mm. The slicer is surely informed by the CAD design, but it slices according to what the printing environment declares. So, it will lay down a layer at Z = 2.00 mm, for example, and the next layer at 2.20 mm. Now, that’s off by either 0.05 mm or 0.15 mm. That is not inaccuracy. Those are limitations of the printing environment.
Using my own example, my printer is very accurate, creating a 20 mm test ciube at 20.00.mm by 20.00 mm by 19.98 mm. I’m using 3 significant figures precision here because that is the precision of my callipers..That same printer will create layers at, for example, at 2.00 mm and at 2.20 mm in Z-height. That’s off by either 0.05 mm or 0.15 mm, though my printer can be very accurate. My cube came out accurately because a 0.20 layer height is a factor of 20 mm.
This is how slicers work. Create a test cube that is modelled at 20.05 by 20.05 by 20.05. Slice it and then look at the sliced preview. You will find that it is sliced to put the final layer at either 20.00 or 20.20 mm, but not at 20.05.
@JonRaymond Would it be possible to move the most recent posts, starting with comment #37, to a separate thread: “Precision of FDM”
Starting with the first comment by Isaac_Citrom, that would be #37.
That way, two completely different discussions won’t get mixed up, and the good comments from Isaac_Citrom will still be preserved. Then we could continue this side discussion there.
