Weird striping effect


#21

All tests I’ve made with Simplify3D and the profile @peopoly posted failed so I wouldn’t know if it’s Cura. But looking at the gcode the model seems fine.


#22

Just an observation: these reports of print artefacts have seemed more common with the recent firmware release. Do you have an older firmware version, either 1.14 or 1.15? If you don’t make any headway with the slicers, try a firmware roll back. If anything, it will rule out that as a source of the issue.


#23

I have the artifacts on parts that I printed with 1.15. Is it possible that earlier firmware versions did not have this?

These lines are parallel to the vat & build plate; the model was tilted slightly:


#24

These parallel lines/artifacts to XY are definitely not random. They run along certain details or support contact points. I had them with 1.15, and with 1.16.

@peopoly How do you mean due to the model? The geometry is pretty clean on all these models, at least shown here in he early posts in this thread. If meshmixer gives the green light, we accept it as being clean. Now if meshmixer is not to be relied upon 100%, let’s start looking at alternatives.

I agree with @ccox here, on the culprit being Cura. I believe his post here is addresses it very well

In the meantime,

Serious 0 - 20000 RPM control from foot pedal, serious torque and Jacob’s chuck.


#25

I believe I can confirm it’s the slicer. I was able to visualize the gcode in a more reliable way and the defects we are noticing are all there. It’s tricky because you see them only at certain angles but they are definitely present.

I played a little with the slicer options but I wasn’t able to remove those artifacts. It must be something the slicer does for some kind of optimization but apparently there’s no way to get rid of them. Slic3r does that too.

We desperately need a dedicated slicer, guys & @peopoly


#26

A proper dedicated slicer would be awesome to have.

I ran the gun model stl through netfabb analysis and got 0 errors. Then I deliberately created bad geo, deleted a ~40 micron triange to make a hole, split a vertex, laminated some faces. Both netfabb and meshmixer found them in the analysis/inspection. So, checking tools work.


#27

Finally you guys are closing in on the culprit! Good work!


#28

Have you tried remeshing the model in meshmixer? Because your wireframe of the model reminds me of pictures I see all the time for game models where people have inverted normals on the face of the polygon. I would be interested in inspecting this model if its not too proprietary,


#29

@matt3o good job on the layer heights! 0.06mm works nice and smooth.

The artifacts though, remain.


#30

yes I tried to remesh it but nothing changes.

The only thing that works is to put the model flat on the bed, which is less than optimal of course.

Anyway I can surely share the model if you guys want to make some tests.

@Thanasis I’m working on it! We can beat this monster! :slight_smile:


#31

The more we narrow it down by elimination of factors, the better. Then maybe a fix for the issue can be found. Poking some cura settings, like “Compensate Wall Overlaps”, “Horizontal Expansion” didn’t change anything, and “Ignore small Z gaps” activated, has a negative effect on other details.

I’m going to try and see if cura handles better a simple object like this, with no continuous topology, only intersecting forms.

Running a dynamesh at 1024+ in ZB and then go for decimation, would give you quickly solid and error free (albeit not the most optimised) topology and a model that on the viewport looks like your dog chewed on it. But it still prints “chewed” edges just fine cause the imperfections are quite smaller than 50-60 um.


#32

I prepared a stress test for the moai (actually the slicer). Look at how bad it is! can’t believe I haven’t noticed this before…


#33

What does the model look like if you open the stl with another program? Look at the mesh in NetFabb or something like that… Seeing those ridges in the solid yellow model above makes me thing they’re part of the file and not a slicing artifact.


#34

your request is legit, but the model is 100% fail proof :slight_smile: The picture above is a representation of the gcode generated


#35

Not a “dirty” export… Huh.
Thanks for humoring me.


#36

@matt3o

can you show how to produce this in Cura. We always have had hard time doing gcode preview. Cura and Slic3r has a lot of FDM optimized functions, including but limited to polygon reductions.
How would this interplayer with the layer heights as if you move from 50um to 60um?


#37

@matt3o
Can you please post a link to your “slicer stress test” stl. I would like to play around with it in cura.


#38

the only software that exasperate the effect is Slic3r. It’s not a 100% representation of what it is going to happen but it gives you a pretty good idea. Unfortunately it doesn’t have g-code preview generated by other software but it looks like the glitches are exactly where are shown by the preview (also by cura).

@nbltsgndpfrdbrms I’ll print it (I haven’t yet) and post the 3D model (just to verify that everything’s fine)


#39

can you measure on screen that it is the same distance between each line as what you seen in the printed object


#40

It is printing. I’ll let you know in about 20 mins :slight_smile: