Hi guys,
I would appreciate your thoughts and comments on the following. Before I go any further I use an Anet A8 with a 0.4 mm nozzle.
I was having some issues with the thickness of internal walls in quite a complex model that I was printing. The walls were coming up over thickness and I finally resolved the issue by reducing the designed thickness of the walls in the 3D model. The new parts I was printing then fitted up perfectly with some existing proprietary parts.
I was having a look to see how the problem had arisen so I printed off some test pieces. These are, if you like, the bottom half of a hollow cube - a thin walled box open at the top. These can be modelled and printed in two different ways: a) by designing as a solid block and printing with 0% infill and no top infill, with the wall thickness set using the shell thickness in the slicer (Cura in Repetier Host). or b) by designing the object as a thin walled block and then printing with 100% infill and no top infill.
In both cases the printing was with 1.75 mm set as the filament diameter, 100% flow, 0.4 mm nozzle diameter and 0.2 mm layer thickness.
The two tests printed as per method a) came out pretty well. The block is 30 mm x 30 mm and I tried 0.8 mm and 1.2 mm as the shell thicknesses in the slicer. The overall dimensions of the 0.8 mm wall thickness block came out at 30.06 mm with an average wall thickness of 0.82 mm (based on 3 readings on each side of the block). The overall dimensions of the 1.2 mm wall thickness block came out at 30.10 mm with an average wall thickness of 1.21 mm.
When I printed the second one again using method b) - Hollow model with designed wall thickness of 1.2 mm and 1.2 mm shell thickness set on the slicer, the printed wall thickness came out at 1.6 mm. The only way I could get the correct thickness was by reducing the thickness of the walls in the model to 1.0 mm and again printing it with shell thickness set to 1.2 mm. Similar results were obtained with the shell thickness set to 0.8 mm but the error wasn't quite as bad.
The first method should, and does, give accurate results on a well calibrated machine. Any ideas why the results from the second method are far from accurate? The model is too complicated to design as a simple solid model with a specific wall thickness and while I can allow for the effects above in the design of the model, it would be much better if I didn't have to.
