Hi Daid,
I'm glad you like it. Perhaps you can use it for inspiration for you own slicer. I'll try the program after I feel Netfabb's under control and I can get good quality prints every time I turn on the machine.
Looking forward to other user's experiences and some print from the Slic3r's gcodes.
http://daid.eu/~daid/IMG_20111107_195330.jpeg
Left is a print with skeinforge, right is with slic3r.
Part of the uglyness is my fault, part Slic3r I think.
I didn't get the start/end gcode to adjusted. So I manually put the M92 in the final GCode. And the UI doesn't remember anything by default, I really suggest you save the current settings on closing the tool, and reload those on start. Just from a useability standpoint.
http://slic3r.org/This is pretty interesting as a quick-calculating alternative to Skeinforge. I tried some slicing with large models and found the calculation to be really quick. Due to the lack of profiles and not being able to make one myself (due to lack of experience) I didn't try to print the gcode.
I just printed a quick test, after inserting some extra code into the start and end manually:
bottom_layer_speed_ratio = 0.3end_gcode = M104 S0 ; turn off temperature\nG28 X0 ; home X axis\nM84 ; disable motors\nM107 ; turn fan offfilament_diameter = 2.89filament_packing_density = 1fill_angle = 45fill_density = 0.4infill_every_layers = 1layer_height = 0.2multiply_distance = 6multiply_x = 1multiply_y = 1nozzle_diameter = 0.5perimeter_feed_rate = 30perimeter_offsets = 3print_center = 100,100print_feed_rate = 100retract_before_travel = 2retract_length = 1retract_restart_extra = 0retract_speed = 40rotate = 0scale = 1skirt_distance = 6skirts = 1solid_layers = 4start_gcode = M92 X78.7402 Y78.7402 E926\nG28 ; home all axes\nM106 ; turn fan ontemperature = 220travel_feed_rate = 200use_relative_e_distances = 0z_offset = 0
The infill path is slow and ineffective: it looks like it's not a continues path, but single diagonal lines with a jump&filament retraction between them, making the extruder motor going back&forth like crazy, plus it's causing some blobbing. the upskin infill is not waterproof (has about 80% density), and is pretty ugly.
other than that, I love the speed and simplicity.
Here's my configuration for Slic3r 0.5.1beta. I get a pretty clean print on a calibration cube. I'd like to test it on more complex objects (like magNEATO for example 
but unfortunatley slic3r doesn't handle that well yet.
bottom_layer_speed_ratio = 0.2 duplicate_distance = 6 duplicate_x = 1 duplicate_y = 1 end_gcode = M104 S0 ; turn off temperature G1 Z+30 ;move Z up a bit and retract filament by 5mm G28 X0 ; home X axis M84 ; disable motors M107 ; turn fan off filament_diameter = 2.94 filament_packing_density = 1 fill_angle = 30 fill_density = 0.4 fill_pattern = archimedeanchords first_layer_height_ratio = 1 infill_every_layers = 1 layer_height = 0.2 no_extrusion = 0 nozzle_diameter = 0.32 perimeter_feed_rate = 30 perimeters = 2 print_center = 100,100 print_feed_rate = 80 retract_before_travel = 2 retract_length = 0 retract_lift = 0 retract_restart_extra = 0 retract_speed = 40 rotate = 0 scale = 1 skirt_distance = 20 skirt_height = 1 skirts = 2 solid_fill_pattern = rectilinear solid_layers = 2 start_gcode = M92 X78.7402 Y78.7402 E926 G28 ; home all axes M106 ; turn fan on temperature = 240 travel_feed_rate = 250 use_relative_e_distances = 0 z_offset = 0
ArunC posted a topic in UltiMaker Cura,
.thumb.jpeg.0b7a05eafc09add17b8338efde5852e9.jpeg)
Recommended Posts
Daid 304
Interesting. It also looks like an easier UI then skeinforge.
I think the following configuration will work:
(save it as config.ini and load it)
As it seems to use the same 'dimension' type of configuration as the new SF40/SF41 you will need a M92 in the start code for the right number of E steps. And at least a sprinter/marlin firmware.
I did a slice test (no print), and it's faster then Skeinforge. It does have the same 'feature' as skeinforge that it hangs during slicing, and only produces console output. It also doesn't remember the last selected path. I guess the developer is using the command line himself.
For my test model:
Skeinforge: 6 minutes 13 seconds (with psyco, and speedup patches found in SF43)
Slic3r: 50 seconds (with the above settings)
SliceAndDaid: 1.8 seconds (Cheating, because doesn't generate GCode yet, doesn't do infills, but does generate an path for the perimeter)
So, it's at least 6 times faster then using Skeinforge. However, it's Perl, so my Java slicer should run circles around it (once it works). It does generate GCode with comments in them, so make sure your GCode sender strips those out or you might run into buffering problems. But it looks like a valid option instead of skeinforge.
Link to post
Share on other sites