On this page of online documentation:
https://docs.blender.org/manual/en/3.6/addons/import_export/paper_model.html#usage
There is a definition of ‘Zero-area faces’, using this line:
‘typically their vertices lie all in line’
Can anyone give me an example of such a geometry, please?
From https://www.blenderbasecamp.com/
What Is The Definition Of A Zero Face Or Zero Edge In Blender?
A zero face refers to a face that falls below a defined threshold in terms of its size. In the case of 3D printing, the printing can only create detail within a certain range. Some for example require parts of the model to be no smaller than 2mm.
Ok, thank you so much for such a detailed answer. So all vertices of the face are co-planed, am I correct?
Hey,
Can you please tell what you mean by saying ‘co-planed’? I just didn’t get it.
From what I understand, zero-area face - is a vert/area/face that is located in a place that is difficult to detect with the naked eye. For en example:
I am facing with that problem too, from time to time, the one thing that helps me a lot is to activate a wireframe mode along with X-Ray tool while being in an edit mode (see screenshot below). This helps me to detect those ‘hidden vertices/edges/faces’ and remove them.
@Debuk, am I right with my conclusion? Just curious.
Cheers,
Sergey
No. A triangle as primitive is per definition always planar.
The idea is pretty simple, its face area is zero. So there are some cases possible. Three vertices in a row (colinear) or 2 or more vertices of a triangle are collapsed.
This 3dprinting size treshold is no real zero area faces. Thats why I simply would forget about that.
@goorman: Well you wont neccessarily find them all that way. They can easily fall togehter pointwise with the vertices of neighboring faces. Depending on the specific type of a distorted face and if its isolated or not there are different ways to get rid of them. For connected faces merge by distance with distance 0 is a good way.
Sure thing, I am not pretending that my method is the best of the best, just decided to tell how I am doing this in my own case 
Anyway, thanks for the explanation
This wasnt meant negative its just that you really will miss cases.
Yeah, I agree. Sometimes I do missing some of them
Just double checked it, but clean up is good for that. It removes isolated verts and collapsed faces aswell. ( degenerate dissolve and delete loose )
Yeah, exactly! Nice explanation of this I might say
Sad there is not buttons like ‘check that model for hidden vertices’ and after that button ‘delete them all!’
No what I don’t understand is the line:
‘vertices lie all in line’
to me, they are all on a straight line, meaning if that line was extended to a surface, all vertices are co-planar, or lying on the same surface. But now I really got the zero area bit, that is vertices do not form a coherent surface (ie. duplicated vertices that do not contribute to form a surface), is this correct?
If we are talking about the mathematical definitions a plane is never spanned by three colinear vertices. An additional degree of freedom is left so that an infinite amount of planes are still possible. Three points that are NOT colinear span a plane.
If you take the a vector from two points and the third one can by reached with a scalar multiplication its colinear.
The video I made for you demoes such a case where at the end there are three points in a row. A face that consists of these three verts has a face area of zero.
Also be aware that a face and plane are not the same. A plane is infinitely large. Its like the difference between a line and a segment
Create a grid. In edit mode, select 4 vertices of one face.
Then, scale face to zero. You just created a zero-area face.
If you do same thing, while constraining scale on one axis, you obtain another type of zero-area face.
Vertices have not been merged into an unique vertex or edge. They are still 4 vertices of a face, that can be individually selected and moved.
There is still a face with an area equal to zero, that is useless to define surface.
It could be replaced by one vertex, one edge or one edge loop.
So that mean all vertices that share the same coordinates (same vector values, location) are the potential zero area faces, but not ‘all lie all in a line’, am I correct?
No. All vertices sharing all coordinates is one possibility of zero-area face corresponding to a point.
The other possibilities are aligned vertices describing a line.
All vertices that are at same point, can be considered as on same line.
If you scale a grid face to zero, but scaling is constrained on X axis : vertices are sharing their coordinates by pairs, looking like an edge, not a point.
If a face is made of first vertex at coordinates (0,0,0), second at (1,0,0), third at (2,0,0), fourth at (3,0,0) : All vertices are aligned. None has same coordinate of another one. But edges between vertices are overlapping. And face has a null area.
Yes, I did just this, scale a whole face of a corner of the grid and 3D-Print reports just that:
thank you very much!
Excellent, the last three coordinates really shown that they are indeed on one line. Thank you so much! But I did this example, took a cube
went into edit mode, added several loop cuts, scaled X, Y to 0, leaving Z alone, but the 3D-Printing reported Non-Flat Faces only, not Zero Faces. I am confused at the definition of Zero-Area Faces now!
The 3D-Print toolbox addon has to detect non flat faces, when two tris assembled as a quad are no co-planar.
That is normal that the addon considers this type of zero area faces as non-flat faces, too.
For a simple case of one face, the addon is considering my zero area face as a non-flat face.
So, maybe, the alignment of vertices is not perfect, they have been slightly moved above Degenerate threshold, by inadvertence.
Maybe, in the way, you created the zero area faces, you encountered a bug with the addon.
