Now that I have a completely stock Rostock MAX V3 I thought I would put the common slicers through their paces as I developed slicing profiles for each. Here are the setup details.
Five slicers, all the most current versions and all running on MacOSX 10.11.6:
- KISSlicer 1.5
- Cura 15.04.6
- MatterSlice bundled with 1.5.3.6894
- Slic3r 1.2.9
- s3d 3.1.0
.2mm layers (all of them)
3 top and bottom shells
2 perimeters printed inside out if available
extrusion width set to 0.5mm (to match nozzle)
30% straight infill
hotend: 195°C
bed: 55°C
bed surface: PEI
first layer print speed: 20 mm/s
all other print speeds (infill, perimeters, etc): 50mm/s
retract: 1mm
retract/advance speed: 25mm/s (except KISS which allowed me to use 50mm/s retract with 25 mm/s advance)
Z lift: 4mm
Travels: 200 mm/s on X, Y and Z
I sliced my BandAid model and saved gcode for each of these slicers. I used OctoPi to print all of them as it eliminates potential jitters and other print issues related to desktop USB control applications. Two of each were printed in SeeMeCNC's black PLA (except the Slic3r gcode was printed 3 times as mentioned below). The BandAid model is relatively simple but has some features to test the slicer. A calibration model of some sort would be ideal but that will have to wait for the book. Here are the results with the best on the left to the worst on the right. Comments below.
KISSlicer: although maybe not as obvious in the photo, the parts were much cleaner, more consistent and had better surface finish on both the perimeters and top and bottom surfaces.
Cura: output some respectable gcode for this model. Interestingly, the bands did seem to be slightly more pronounced than with other slicers on both test prints. Otherwise, it was quite clean.
MatterSlice: the primary issue I had with MatterSlice was that even at the same print speeds (and I verified these in the gcode) the print came out a bit glossier than the others indicating a slower print speed for the perimeters.
s3d: had some artifacts, particularly the "warts" that many see. In addition to the set of defaults I used for the other slicers I did attempt to tweak the s3d parameters to improve things. I was able to get slightly better prints.
Slic3r: is an odd one. You might not be able to see it in the photo but the perimeters were horrible. It looked like the printer might have had a loose belt or arm or joint but that was not the case. In fact, immediately after the first Slic3r print (that was horrible) I printed the KISS code and it was perfect, then the Slic3r code and it was horrible. I scrutinized the gcode and all of the parameters were exactly how I configured them (as described above). I ended up printing 3 just to make sure. I don't understand this at all. I attempted to tweak the slicing parameters to improve things with very marginal success.
Since my V3 is absolutely stock, if you have a stock V3 you can print my gcode files for yourself to test (all except the s3d gcode since their license prevents users from sharing code). I've attached the files here along with the BandAid STL in the zip file.
If you scrutinize the gcode and see anything amiss, let me know and I can rerun the print. I tried to be as consistent as possible.