Are Slicers Accurate Enough?

[CrazyIvan]

I'm sure we've all experienced printed parts not fitting, and the pragmatic solution is to do test prints of critical sections and adjust clearances in the design so that they do ultimately fit.

So, when calibration cubes come out just right (within a few hundredths of a mm), why don't things fit? For one thing, holes close up a bit no matter what, and that's one reason default settings have the flow rate substantially less than 100% (100% being the extrusion amount required to exactly fill the volume of the drawn line).

The nozzle cannot extrude a perfect rectangular cross-section, so there are always voids, but if the flow is 100% it results in a wider line than intended, and it overhangs the design edge of the part, but reduce the flow enough to counteract this and you have a weak print. As I have said many times, FDM printing is about balancing contradictory requirements.

However, what I didn't realise is that the slicer might not be doing a very good job in the first place. I think this has not attracted much attention because of all the other errors in the process, but I came across a video where the slicer output was reconstructed into a 3D mesh, and the sliced mesh compared with the original design mesh – finding errors of up to 0.5mm (which is irrespective of printing inaccuracies, those need adding on top)!



We'll still need pragmatism, but you would think we could trust the tools...

[designforlife]

Interesting! We often blame dimensional inaccuracies entirely on printer calibration, material shrinkage, or design clearances. We forget that the slicer itself is rarely scrutinized as a source of geometric error.