2
$\begingroup$

I am new at this and maybe my model is not the best, I adapted it for another one actually. Can you tell me why I can't get the holes printed? I already checked the faces and they are all in the correct orientation (I think)

What happens is that I start printing with the holes facing down and they are not printed at all. I never let it keep going for long but it seems to be completely filled inside.

You can check the file here.

enter image description here

enter image description here

$\endgroup$
10
  • $\begingroup$ Please consider to edit your post with the original file you modified and also the program with which you created the modifications. If you are able, use the original program to address the specific areas of the model indicated by the red lines. $\endgroup$
    – fred_dot_u
    Apr 8, 2022 at 15:48
  • $\begingroup$ I see inverted faces.. your cylinder is it inside out $\endgroup$
    – Trish
    Apr 8, 2022 at 17:52
  • $\begingroup$ @Trish, I don't understand what you mean by inside out $\endgroup$ Apr 8, 2022 at 19:42
  • $\begingroup$ @PatriciaMartins, STL files are a bunch of triangles, which have an inside and an outside. If a triangle represents a surface viewable from the outside, it is an outside triangle. If a triangle is misplaced, the inside will be visible. In Meshmixer, the result is a zebra-striped surface, none of which are visible in the image I created/attached to the answer. $\endgroup$
    – fred_dot_u
    Apr 9, 2022 at 0:16
  • $\begingroup$ Additionally, when I performed a plane cut in Meshmixer, perpendicular to the axis of rotation, the internal surfaces became visible, and were zebra-striped, which tells me that the insides are on the inside, where they belong. $\endgroup$
    – fred_dot_u
    Apr 9, 2022 at 0:18

2 Answers 2

2
$\begingroup$

The problem is internal geomoetry

The body you modeled consists of a non-manifold shell: There exists a fully enclosed shell on the inside of the item that tries to define an "outside" of the body. In the following picture, I have hidden part of the geometry to better show the problematic internal surfaces in orange:

internal Geometry

Automatic processes such as Meshmixer or Windows 10 3D builder interpret such an internal, one-sided open cylinder as "This probably is missing a surface on both ends". This solution leads to two intersecting and manifold shells - a cylinder overlapping the drilled holes and the body with the drilled holes - which then promptly get treated with a boolean union... and voila! No more holes. Or even no more outer shell as the easiest solution is to just stitch that lower surface and discard the rest. This is what happens when Meshmixer does just that: you are left with the cone half and some inverted artifact areas.

Meshmixer automatic repair removes half the model leaving a few inverted artifacts

So the best solution is to ensure the parts don't contain such volumes encased by a non-manifold surface in the first place. Due to the nature of the part, in this case, it is rather simple: Simply removing the circle of vertices that spans up both the plane, as well as the cylinder, marked orange results in all internal surfaces getting removed. Note that due to the orange parts sharing (at least partially) vertices with the wanted outside, this has to be done manually. Would both surfaces share no vertex, a simple "separate shells" operation could result in a very quick way to remove offending structures.

Without internal geometry

Without the internal geometry, the model gets interpreted correctly - the mere presence of such superfluous internal geometry makes the slicer believe that some surfaces are inverted or missing, and thus need to be inverted or stitched - and the solution to the slicing is... utter mess.

Cura rendition of the defect and cured model Slicing solutions, before and after removal of internal geometry

$\endgroup$
3
$\begingroup$

The overall appearance is that the normals are "normal," that you have no reversed facets, but there are discontinuities within the model that Meshmixer and Netfabb show as failure points. Windows 10 3DBuilder also attempts a repair which fills in the holes.

meshmixer failure indicators

The Meshmixer capture image shows red lines and points at the flaws in the model. Both above programs fill the faces, which works fine on the cylinder, but fills in the plane where the holes reside, as well as removes the internal holes/cylinders, preventing a simple plane cut repair.

Additional examination of the original model shows an internal cylinder formed axially on the end face red warning markers in the image above. I used Rhino3D v6 to slip inside, select the cylinder and remove it. Because the cylinder is "inside" the overall model, there's no inside face and outside face, causing the software to glitch.

On the red line along the circumference of the cylinder, there's an internal disk/disc with an internal diameter to match the previously removed internal cylinder. As it also resides within the overall model, the same trouble applies: no true inside/outside surface for the software to comprehend.

Once these were removed, a problematic set of errors appeared. I'm working on that. Work completed. Windows 10 3DBuilder has a pretty amazing repair facility, once the deep stuff is cleaned away. The end result passes the Meshmixer Inspector test and I suspect will work for you. The cylinder appears to be 37 mm in diameter (about an inch and a half) which is rather tiny, but with the flaws repaired, will scale up just fine.

It loaded into Simplify3D slicer with no errors and appears will print nicely, although with support required along the beveled portion of the cylinder.

$\endgroup$
2
  • $\begingroup$ Could you please post the last paragraph as a comment on the question? (And remove it from the answer) $\endgroup$
    – 0scar
    Apr 8, 2022 at 15:35
  • $\begingroup$ Your repair too just removed the internal geometry. The twice "normals this way" in a row results in the slicer interpreting one side as "this MUST be inside out or is lacking a surface" $\endgroup$
    – Trish
    Apr 9, 2022 at 1:44

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .