My latest adventure in #VelocityPainting. This is a 0.5mm walled vase-print (so one continuous spiral), constant 210C, 0.2mm layer height, 20cm tall. Mostly printed at 13mm/s, with the line-drawing printed at 35mm/s. M3D Fuchsia translucent PLA.
This is post-processing, so the vase (own design at Curvaceous Vase by markwheadon - Thingiverse) is unmodified and sliced pretty much as normal, with the rose drawing applied to the GCODE file before printing using a perl script to split vectors where they cross pattern boundaries and then tweak the print speed according to the pattern. The rose pattern goes through the vase like letters through a stick of Blackpool Rock, so there’s the same pattern but reversed on the rear side of the vase. Beautiful.
This is absolutely amazing. Can’t wait to see it released! Have you tried playing with temperature instead of speed? What are the results/cons/pros/differences? I’m imagining that increasing the speed could also produce under extrusion, no?
@Shai_Schechter It does seem to cause under extrusion as well, yes. Although I don’t fully understand why as the amount of filament needed shouldn’t change with speed?
Just a question on the under extrusion… Your srcipt adjusts speed on the travel… But does it adjust speed on the extruder feed?
Completely, utterly, unbelievable, fantasic work… BTW.
@Matt_Barth No it doesn’t. I assumed that, say, G1 X<…> Y<…> Z<…> E extruded the filament at the correct speed to evenly spread over the length of the move. If that’s not the case then I obviously need to read up on it!
@Mark_Wheadon Under extrusion is a problem I am encountering. I think it is related to overall speed of the layer as well as size of transition between speeds. In the example below I noticed that the first 50 layers were not affected by underextrusion, but those layers the also mostly made up of a different intermediate speed. missing/deleted image from Google+
This is just my thought. And I may be way off, because this is math…and it hurt my brain…but the extrution in gcode is given in length, so if you move faster you will still put out the same length, but move x times faster. Giving you a thinner line. Like if you are driving at 5mph , and drop your oil pan, the trail you leave is wide, now do it at 60mph the line is thin.
I wonder if the under extrusion is a lag in the response of the Bowden system to the rapid change in speed? Does a direct drive provide the same effect?
@Matt_Barth I don’t think so, because the system is supposed to drop amount of oil in that distance, so the pouring rate is changed to match the speed change – move fast and pour fast, move slow and pour slow, either way you’ve poured the same amount of oil over that distance.
@Jeff_Parish The thing is, the filament is always under compression, so there won’t be much slop – you get the slop with retraction because when the extruder pulls rather than pushing, the push-fits move and the filament slops to the other edge of the Bowden (inner edge of the curve rather than the outer edge). But with this changing velocity, it’s still always in the same direction, so I wouldn’t expect much slop. The thinning also persists over at least a couple of centimetres – so it’s not a quick transient effect.
You should calculate what your max feed rate looks like. If you say “extrude 100mm”, at feedrate X, and the extruder nozzle itself can’t take that pressure load, what I believe will happen is that the pressure will build up in the nozzle a little bit more than usual and then let out with a minor delay, causing less filament to be extruded. My hypothesis would be that under extrusion would show up more at the start of the faster segment and you might have a small artifact at the end depending on how much pressure was remaining and how quickly it equalized out.
@Justin_Nesselrotte Typically I slice with the guide speed set to something like 40 to 60mm/s depending on how much quality matters, so 35 should be undemanding.
