Weird striping effect


#1

No matter what I try I can’t get this texture to print correctly. It’s basically a dotted texture on a flat surface (holes are about 1mm wide).

Apart from the layering (that is understood) each row of holes has two deep stripes on top and bottom. I tried by changing orientation, angle, supports… the problem is stronger in certain orientations but it never actually go away.

Any idea?


Accuracy at Any Size Using Blender's Warp Modifier
Benchmark for level of detail
Consistent inconsistency in Z axis (thinking it's Z wobble)
Advance Calibration, some doubts
Printed object with deep stepped lines
#2

Have you looked at the gcode in Cura? The software makes certain adjustments to the model to optimize tool paths and correct for limitations in the printer, it is possible that the regular dimpling is throwing off the algorithm.

If the line is not in the gcode, it may also be introduced by the Moai firmware, which is making its own adjustments to the gcode.

As unsatisfying as it sounds, you might try scaling the pattern, i.e. change the dot or row pitch, to see characterize artefact and then correct.

I recognize that you have tried different orientations, but have you tried printing the model flat, e.g. 90degrees the the build plate axis? The solution may be as simple as reducing the incident angle of the pattern to the build axis.


#3

thanks for your reply. I checked the gcode and I can’t see any trace of those lines. I’m now trying with another resin but I doubt it would make any difference.

I feel like it has something to do with resin remaining trapped between the holes… But honestly I printed very complicated models without problem. Since it’s a very symmetric texture it might trigger some weird optimization like you said @Paradoxical_Cat to fix something it thinks is a defect.

I dunno, maybe @peopoly can give me some pointers.


#4

What was the orientation of the model? I don’t think it is the resin. I am guessing that the software is some interpreting the space between the rows of holes as raised. Have you.looked at the model in Meshmixer? Depending on how you assembled the elements, there may be a hidden manifold/object there. To my eye, it appears that it was carried over in a cut and paste operation.


#5

I tried to print it in every possible orientation except flat down.

I checked with meshlab and everything seems fine. Consider that it’s a procedural surface, the holes are not made one-by-one (or copy-pasted). Also it’s an extrusion (not a boolean that can create problems).

Now, that being said I opened the model on meshmixer and it is being rendered with some artifacts (not just in the incriminated area), but that would be the only software showing issues. Also the repair tools say that everything’s fine.

Why would the effect become more or less apparent by changing the orientation if it were an issue with the 3D model?


#6

I have had a similar effect with some prints, seems like there is a tendency to create lines like that at specific details. Always getting a line parallel to the XY plane. It is like the laser is doing double runs over the same path, at the height of a specific detail -at the top and bottom of each circle in your case- and it looks like over curing.

Can you share a wireframe of the triangulated stl from meshmixer?

Meshmixer loads the model with normals set to soft, so it will show some areas darker as if they have artifacts but it should clear out if you choose the option to make all normals hard. If the inspector didn’t point out any errors, the geo is good to go.

I believe it’s a slicing issue, whether from cura or from the way firmware handles the gcode.


#7

Every time you change orientation the slicer recuts the model and generates totally new tool paths. Ideally, those toolpaths will be exactly the same other than relative rotation around an axis. However, it is not an ideal world and slicers will sometimes generate different gcode for the same model in the same orientation.

It seems, from the photo, that the holes are actual part of a recessed feature with the left side holes as the anchors for the feature. The question is whether the recessed portion is being generated in Cura or in the Moai. Have you tried to print the model on a different printer? However, Cura for Moai does not handle models the same way “normal” Cura does, but how deep into the code that goes, i.e. into all modules, is beyond my knowledge.


#8

ah-ha! So we are into something here!

You are right changing the rendering mode makes the model look good on Meshmixer

This is the wireframe mode of the model.

Looking at your pictures it doesn’t seem something related to my texture, but any slightly recessed geometry would do that.

I don’t think it’s the slicer because inspecting the gcode the surface is flat. So either moai firmware or something mechanical.

I believe we need @peopoly intervention at this point.


#9

so what you are saying is that you get the highlighted band no matter the orientation? and perhaps not related to layer heights?


#10

I tried to rotate the piece in different orientations. The result is not always the same but the problem is always there. Sometimes the stripes are longer (like the picture I posted above) sometimes they are a little shorter, but so far I’ve never been able to completely eliminate them.

Honestly I only tried at 50um so far.


#11

The geo looks right, it is nicely optimized by your CAD package, maybe Cura has an issue with the vertices where the triangles for each hole meet, high edge density there and maybe it’s tricked into drawing more paths. It’s not z wobble for sure, cause the artifact / line doesn’t spread to the rest of the object.

When you rotate and change orientation the artifact lines follow the XY plane, right?

Can you also try a similar, with more evenly spread out polygons version of the pattern and see if it makes a difference? Something like this

CAD will probably not let you have this, it’s going to want to triangulate it more economically. What you can do for a quick test is to make the part hollow there and add the D shaped polygon mesh with the holes created from a poly modeling package. You don’t need to merge them, as long as the volumes overlap the slicer won’t mind. Meshmixer will give you a pink warning but that just means you have 2 solids, it will slice and print just fine.


#12

The problem is worse than I thought actually. Today printing some tests I noticed that the same kind of error happens on the supports. When the feature is recessed the error is recessed too, when the feature is protruding (like the supports) the error is protruding too… I don’t remember this happening before, is this something introduced with the new firmware?


#13

I got this same artifacting with 1.15, seems to be doing it where the supports are touching and where there is some particular detail, like the edges of circular forms. At first I thought that it was a case of slicer or firmware getting mixed up with the intersecting volume of support tip and model geometry. So I took the same model, this time fused in the supports and re-triangulated, rechecked in mmixer, came out clean, sliced, printed and the same artifact appeared.


#14

this is pretty bad, can’t believe this is coming out just now


#15

50um Jun18 profile,
modified 0.0679 Layer height
Temp 30 C

Banding, indentations and lines. This is not random, or some wobble, the artifact follows specific details and will always appear along the details even if you change rotation or heighten supports and increase distance from build plate.


#16

I’ve noticed these anomalies on models before. There’s another thread from a few months ago about them, too. The ghost lines don’t appear on all models, just hard-surface mechanical models, and they seem to occur in the worst spots that are hard to get to with sandpaper or a file. I can only imagine they are caused by the laser not being in the exact right spot as it traces around a bump or detail and then is supposed to continue on in a straight line to match up with the previous layers that did not have that feature present. I wonder if really slowing down the print speed would help with the positioning? I know the artifacts occur with the slower 55-laser profiles, and I don’t think the laser power can be dropped much below that with the standard resins. Speaking of that, it could be a problem introduced by the resin itself. Has anyone seen the problem occur with another brand of resin?


#17

it is best if you got some pictures of different orientation of the same print.

There are two key things to look for so we don’t mix all different factors

  1. are anomalies parallel to XY plane.
  2. are the random or always in a fixed position or fixed layer counts.

In general, if it is not random and doesn’t change post when orientation is changed, then it is likely due to model
if it is random and parallel to X/Y, it can be mechanical, electrical or software. it is very unlikely to be resin.
If it is fixed position and parallel to X/Y, then it is often mechanical.

some users print at that 55 profile but not many. It may not be a good profile for large print. @Joseph


#18

Simplified3D released a Moai profile, it is not fully tested by us yet but we could print with it. It does something funny with the initial layer and then is fine after 3 layers. Those of you have have S3d could give it a try and compare. Do a small test run before dropping any large model into it

S3D profile V3 by Simplified3D
https://drive.google.com/open?id=13CEq3yAY6QoBgLApEcNmHSH4r4O9w1T_


#19

I wouldn’t say it’s random, it’s actually pretty predictable and it is indeed parallel to X/Y. It is more obvious on mechanical/flat models but it may also happen on more organic stuff (like in the moai model I printed above).

I’m now printing on Simplify3d, let’s see if it’s Cura (but I don’t think so).


#20

And the lines corresponds to the start or end of other details on the surface of the model. I’d bet money on it being a Cura bug.