To minimize waste, at least less waste than a whole transition tower, could you put a wiper on the left side of the X axis and have it extrude into free space, but off the side into a cup or something, and then drag across the wiper back to the model?
I also saw there’s options to do some of the transitioning in the infill, and some other trick to minimize.
Did you mount it near the toolhead? You might be able to hang the splicer on the back of the X axis which would really shorten the amount of filament needing to be wasted.
Yeah I agree the transition tower looks wasteful. I was hoping the transition could have perhaps happened with the fill alternatively I need to figure out the setting which simply says to the Palette make X filament and the printer extrude X. I’m sure it can probably get with extruding about 40mm - 70 mm from the printer but not sure how much filament that would actually be. At that point you can see the colours blending but once it gets to the model its pretty sharp.
I’ll keep researching but once I get it tuned I’ll repost my settings as a ballpark.
I’ll be watching this, I might get one of those. I think it’d be easy enough to mount a wiper on the X axis so that every time the head goes left of X=0 it drags across a thin metal scraper. That would completely eliminate the need for a transition tower.
No not yet. The only calibration was the print head against the heat bed but I’ll look at all the other recommendations.
I’ve done a few prints with the Luban snapmaker software which have been awesome including a new functional hinge and vase but this is the first print incorporating the Canvas Hub S and Palette 2S Pro…
I think the extruder calibration is going to be a big impact as the machine under extrudes out of the box. The linear advance will also help with the ghosting, and im sure both of these are being augmented by the introduction of additional colors.
I did some tweaking of the model and ran the calibration etc, wasn’t much that needed changed. The roughness on the visor was actually an error on the model itself the visor wasn’t as large as the rim, so the roughness was scaffold at 15% fill. I printed this upside down this time (high speed print) and its nice and smooth now but I’m still puzzled by the colours.
I made a pretty simple calibration print to test all 4 channels of to determine how precise the print was going to be. Results were quite interesting I sense I need to calibrate the Palette 2S somehow…
On the X & Y it looks fairly OK but there is the occasional completely wrong colour, its like a skew effect on the initial layers.
Was this print from right to left, where the grey bled into the blue?
I don’t know about the software configuration - how does it know how much filament to “waste” so that it’s ready to go for the next color. It looks like it’s not enough sometimes. Is it something you input in inches or mm, like there are 300mm between the palette and the toolhead?
No this one was actually more intelligent although “questionable structural integrity”, whilst respecting the z-order, it prints the same colour.
The quantity of filament to “waste” is specified by the transition parameters in canvas. I think these were initially set of 130 mm which I thought was super wasteful after my first test, so I used custom transitions between the different materials. So basically if there are strong colours like the black and silver I would give more transition to the weaker colours. 70 mm to 90 mm. AFAIK I understand this is how much is needed to empty the bowden tube of the previous material to the new material, to reduce the blend effect. I wonder if I need to factor in the bowden tube length?
E.g. If I need to print say 10 mm worth of filament A, transitioning to filament B then printing 20 mm worth of B, before transitioning to filament C and printing say 5 mm of C. Ignoring a constant flow rate (which I’ve calibrated). The filament which is being output from the Palette 2S is approx:
+ 10mm of filament A
+ X mm of filament B (for transition) + 20 mm of filament B
+ Y mm of filament C (for transition) + 5 mm of filament C
Now I wonder if I should consider increasing the length of the transition to be greater than the bowden tube? I didn’t think that really mattered as I was going from the tip of the extruder as a fixed point, its not like its a “blended colour” on the output model. However I do see this on the transition tower. It just look like where the print expects to be the output filament isn’t quite in sync.
I’ll try another experiment similar setup, I’ll reposition the camera and change the timelapse. This time I’ll try to keep it really simple 4 cubes each 1 specific colour.
I did noticed a firmware setting on the canvas, under the printer profile…which I just left as is.
Almost successfully printed 4 cubes in with the correct colour coding with no colour bleeding however I must have suffered a splice join failure or something because printer was “printing dry” . Room for improvement I think on the Palette that it cannot detect this sort of failure, perhaps a rotation detector on the outbound filament.
Anyway latest results looking pretty good, jam aside.
Extrusion:
Extrusion width: 0.4 ← was 0.42
Extrusion multiplier: 100 ← was 103
Gap fill: off <— This I believe was causing the stripe effect on the visor (black n blue)
Retraction distance 0.4
Transition: 130mm (but I still think there is too much waste the transition wall is large than 4 times the thing I’m printing…
I am really digging this, you’re making it happen!
