Thin wall anomaly ?

[lidgaca 0]

Hi

 

I'm using Freecad to design a robot - currently making a wheel hub. I export the .stl file into cura and I currently have cura 4.12.1. I have noticed a weird anomaly with the slicing of thin walls - wondered if it was generally known, or if I need to report it as a bug.

 

The hub unit has some deep square section slots for captive nuts. The slot dimensions are 7.3  x 20 deep x 2.5mm (printer x, y, z), and the 3 slots are arranged at 120 degrees to on another..

 

If I print with supports, extracting the support from the slot is *very* difficult - the depth of the thing, coupled with the very narrow entrance make it basically impossible. I have printed the hub with a support blocker, covering the slots, and that's OK, but the top (unsupported) part of the slot tends to sag and makes inserting the captive nut into the slot either very difficult, or impossible depending ...

 

So I added my own thin section supports in Freecad with the intention of making a support structure that was strong enough to support the sagging roof, but with enough space around it that I could get a small widget inside to break it free. The support walls are 0.4 mm thick, as is my print nozzle. I set the 'print thin wall' option, and the results are shown below.

 

The slot at about 11 o'clock seems fine. The two support walls are clearly printed and the filament flow is seen to be up and down the support wall. Note that this slot is orthogonal to the printer axes.

 

However the other two slots exhibit stranger behaviour. Take the slot at 07:00 ... It is 120 deg to the printer axes, ... but the filament flow is seen to be a number of short segments parallel to the print axes. In fact what is happening for the two non-orthogonal slots is that the slicer prints one line of filament actually along the wall, and then the next slice is printed in these short segments. It alternates up the structure. At the point I'm showing, the slot at 07:00 shows the 'short orthogonal segments' and the slot at 02:00 shows the 'flow along the wall'.

 

I hope that this rather complicated explanation is understood 🙂

 

I have yet to attempt to print this so I don't know if this is will be a problem or not, but I thought I'd open this up for discussion.

 

 

Any questions, and I'll try to expand 🙂

 

 

-- Chris

 

 

[gr5 1,837]

There are rounding errors both in freecad and in cura.  You need to either make those walls 0.41mm or change your line width in cura to slightly smaller (by about 0.01mm).

 

A 0.4mm nozzle prints reasonably well from 0.33mm to 0.48mm - really it can print beyond that - for example from 0.3mm to 0.6

 

 

[lidgaca 0]

I don't know if that really addresses the problem but I will try it.

 

The issue seems to be why does cura happily print one thin wall (the one which is orthogonal to the printer axes) as repeated single lines, but prints the other thin walls as (short segs one slice / single line the next slice) ? I can't see why it doesn't just print repeated single lines for all the slices ...

 

That seems strange behaviour ...

 

 

-- Chris

 

[gr5 1,837]

Oh.  Okay I was confused by the picture because I can't rotate it in 3D.  I thought it was a dashed wall.  Now I see it's diagonal skin.

 

Also I realize now that it's yellow.  Not the expected red.

 

I'd play around with nozzle size just to see what happens.  Set nozzle to 0.15mm just to see if you get red passes.

 

Yellow is "skin".  Not wall.  Not infill (orange seems to be infill).  So to get ready for a wall above, at some point it goes - hey there is a wall above me so I better print some skin here to support it.  But then on the next layer it's like - oh - the wall is too thin.  But "hey, there's something above me.  So I better print some skin to support it".  At least that's my guess as to what is going on.

 

Basically your walls are too thin.  I usually make my walls 0.9mm to be sure they will print.  Even if they are just breakaway support walls.  I might do dashed walls for 2 layers then solid walls and then dashed at the top so they break away easy.

[lidgaca 0]

I have to admit it was very complicated to describe ! So thanks for persevering. Probably not helped by me using the word 'wall' out of context for cura 😞

 

Yeah, I just thought it was an odd result ... I haven't tried to print it yet so I don't know what will happen.

 

I will try and fiddle with wall thickness at some point, and will update this thread if I find anything  significant.

 

Thanks again

 

 

-- Chris

 

[gr5 1,837]

It's MUCH easier to fiddle with line width.  Just to learn.  Only then go back and fiddle with your cad model.

[lidgaca 0]

Just to finish this off I simplified my model so that I could play about with it and do some tests ...

 

You are right, the key part is that with a 0.4 mm thick thing and a 0.4 mm nozzle, the thing is printed as skin in yellow, and then exhibits the strange behaviour. If the thing is thicker than 0.4 mm then the thing is printed as wall in red, and the printer nozzle will move along the structure in the expected way.

 

The illustration below shows two thin structures, the LHS one is 0.4 mm, the RHS one is 0.45 mm thick. The rule seems to be less than OR EQUAL TO nozzle width ==> skin, greater than nozzle width ==> wall.