Dancing around the support fill

[Captain Starfish]

Finally, I think I'm getting my RMax tuned to the point where it's doing mostly what I want as far as going to a certain point when its told to and squirting goop.

Now it's time for a little optimisation, I think.

Presently using slic3r but occasionally jumping to the free version of kisslicer.

Test model attached. It's pretty much a quarter of a sphere that opens into half a cylinder. This is a mould for half the nose for a "shot" - pretty much a torpedo made of lead and that has an eyebolt sticking out of the back of it. When we get to a deeper dive site location (GPS marked) we drop the shot, it tows a line down with a float on the other end. Because it's heavy and streamlined it drops to the site fast, and is less likely to get dragged off the mark by any currents in the water.

The model is pretty much a 0.7mm walled shell with a notch at the open end for clipping the body section into. I expected a smooth double pass around the shell and "swimming lanes" for support. Not at all what I got and not sure why.

Anyway, it printed but there was some crazy head dancing going on. I tried with kisslicer but don't really like the way it slams the carriage about the place and there were heaps of fiddly little jump, squirt, retract, jump somewhere else
sequences which just did my head in.

So I tried again with Slic3r, and watched. Perimeter / shell went ok after I tweaked the model a bit. But two things went funny with the support. First, selecting a 45º overhang I did not expect a 100% support fill which was what I got. Annoying because it adds time and wastes material. Second, instead of just marching up and down the lanes for the support it would go up/next lane/back then jump to some position and back again before continuing. Halfway across the layer it would choose another one of these arbitrary points to jump to. No idea what's going on there unless maybe it's a rounding error and a "no perimeter crossing" setting making Slic3r get all excited and want to dance.

Is this kind of thing normal? Or should I be doing something to settle it down?

[Batteau62]

Can you show us your slicer settings? I use Kisslicer more than slic3er. Curious how many perimeters you used in Kisslicer? And why support? This looks like it should print without?

[Captain Starfish]

What's the best way to show all the settings? Screen grab or config file?

Kisslicer isn't doing the support dance, but everything I've printed on it has been violent compared to slic3r.

[geneb]

All the settings appear in the comments at the beginning of the file that Slic3r creates. (I _think_).

g.

[Captain Starfish]

Of course. Thanks Gene

; layer_height = .2
; perimeters = 3
; top_solid_layers = 3
; bottom_solid_layers = 3
; fill_density = 0.2
; perimeter_speed = 30
; infill_speed = 35
; travel_speed = 400
; nozzle_diameter = 0.35
; filament_diameter = 1.67
; extrusion_multiplier = 1
; perimeters extrusion width = 0.52mm
; infill extrusion width = 0.52mm
; solid infill extrusion width = 0.52mm
; top infill extrusion width = 0.52mm
; support material extrusion width = 0.52mm

Example of the path here: [img]http://www.simonlockwood.net/linky/3dp/dodgyshot.jpg[/img]

It's pretty, but it's slow.

Ended up trying the mid-section with Cura and Kisslicer. Cura seemed ok, Kisslicer still bashing the head around.

[Batteau62]

"; nozzle_diameter = 0.35"
I think your extrusion widths are off. If my memory(getting bad!) serves, extrusion width is like 110-120% of the nozzle size. So a .35 nozzle would be around, .385. I don't know if that will cure your problems, but I would check it.

[Captain Starfish]

I didn't set the extrusion width, as far as I can tell.

I did the thin-wall test cube, though, with each material so I could set the extrusion multiplier and I'm pretty much bang on 0.52mm wall thickness.

[Batteau62]

Weird? Your settings say "nozzle diameter=0.35"? It sounds like you have a .5 nozzle. Is that what you have on the hotend? If your single wall cubes are coming out as .52, that would be hard to get out of a .35 nozzle. Not sure what that would do in the slice?

[Captain Starfish]

Hmm

Now I'm confuzzulated. I tell Slic3r I have 1.67mm filament going into a 0.35mm nozzle and it's spitting out 0.52mm wall thickness when I try to run the 0.4mm thin wall cube. I expect this is maybe a double perimeter with overlap. Regardless: extrusion width measured is extrusion width expected and extrusion width ain't my query, it's all that pesky jumping around for what should be a simple curve.

[Batteau62]

My theory/guess is that having the nozzle diameter not set, to what it actually is, causes the math to be off in generating paths. Especially curves and circles. I remember a post by polygonhell about how arcs are actually very small straight lines. Maybe the slicer is "confused" and is gererating erratic paths, causing all the "pesky jumping"? Just my shot in the dark here. Chime in guys!

[geneb]

You should manually set the extrusion width to be about 10% more than your nozzle diameter.

g.

[Captain Starfish]

How? Just by setting the "default extrusion width"?

[dpmacri]

How? Just by setting the "default extrusion width"?

In the "Print Settings", "Advanced" tab there are 7 "Extrusion Width" settings. For a .5 nozzle, set all but the "First Layer" to 0.55. I usually leave the "First Layer" setting at 200%. For my 0.4mm nozzle, I set the others to 0.44

[Batteau62]

For Slic3r- Yes, set all seven manually in the "advanced tab" You can also express them as percentages computed over layer height.
In Kisslicer-style tab-There is "extrusion width[mm]" and "infill extrusion width[mm]" And "number of loops" is perimeters.
In Cura-The "advanced" tab has nozzle size. Cura uses this to calculate perimeters and extrusion widths based on your settings for "shell thickness" in the "basic tab".
And all slicers rely on accurate measurement of the filament.
Hope this helps

[Captain Starfish]

Thanks very much folks. I'll give it a try next print.