Hi,
After a few months of practice, I just discovered today that Groups approved as" shiny solid groups" by Solid Inspector2 are NOT recognised as such by Sketchup !!!
Yes I used to regularly check my construction of groups with SI2, with approuval, but often failed using Union /Inter/Split tools with “shiny” groups and didn’t understand why?..
Can someone tell me what is(are?) the possible difference(s) beetween the two jugements given ???
and how to detect and correct these hiden detection failures?
Thanks
(using SU Make 2017)
There are some cases where Solid Inspector2 will identify soething as solid but SketchUp won’t. Share one of your models that shows this and we can help you identify the problem area.
I don’t know why Solid Inspector2 doesn’t pick this stuff up but there are all kinds of internal faces. Hide the back face and you can see them. The confusion might be due to the overlapping faces in that area. You can see the Z-fighting.
You have a lot of linear-detailing around windows [front & back] which is actually drawn as a ‘flat’ outline, and so it’s coplanar with the wall faces.
If I edit the main container, select-all and use a coplanar removal tool, then they are deleted, and then the container then reports as a ‘solid’ in SketchUp.
If you want the detailing try to PushPull all of it so it is not coplanar and try again…
Even my SolidSolver isn’t great at fixing it without this kind of manual fix first…
No in fact the blue faces you show here are not internal !
But due to an intentional vertical shift when I exploded these ‘curved door structures’ into their correspondind wall insets : the reason is that SU was enable to propose a correct global intern/extern faces solution to this ‘explode’ process !.. I tried a lot without success .
Finaly I resigned to this work-around: scale down in the Z axis the included doors group to reduce those curved internal faces that seemed to create the problem…(making them external by the way)
NB : what Zfighting really means (sorry I am french…)?
The plane linear detailing you suggest, as well the gap (green axis) mentioned by DaveR, are intended to the next steps : push pull of a global face of this group to stick on it a new group representing the stony tructure-surface of the building…see this more recent .skp reflecting the current state of the project
I am not sure that my methodology is the right one : I wanted tho ‘explode’ the stony structure upon the base structure, to stick/glue them together ; but SU produces incoherent Internal/ external faces at the end…. That’s the resaon why I am posting to you, in fact , as I discovered yesterday this pb of ‘false-true Solid’, which might be the clue ? collage AV_0.skp (7.1 MB)
OK TIG, I used Cleanup3 to to clean all coplanar edges on ‘AV 3D Extrudé’ group, and yes, it is declared as Solid Group by SU now!
BUT I did the same on the other group ‘new revet …’, and it still is NOT recognized as Solid by SU (but still shiny by Solid Inspector2) !
I dont know what’s wrong with it???
and if I explode this ‘new revet’ group into ‘AV 3D extrudé’ I obtain a final group, whose stony faces need all to be wrongly reversed, to be accepted as Solid group by SolidInspector !
What is the problem?
Ok TIG, I used Cleanup3 to to clean all coplanar edges on ‘AV 3D Extrudé’ group, and yes, it is declared as Solid Group by SU now! BUT I did the same on the other group ‘new revet …’, and it still is NOT recognized as Solid by SU (but still shiny by Solid Inspector2) !
I dont know what’s wrong with it???
and if I explode this ‘new revet’ group into ‘AV 3D extrudé’ I obtain a final group, whose stony faces need all to be wrongly reversed, to be accepted as Solid group by SolidInspector !
What is the problem?
Yea, if there are overlapping faces it can throw off how Solid Inspector 2 analyse the mesh.
Solid Inspector 1, which uses the same definition of a solid as SketchUp do might be better to use in this particular model. It’s been on my todo list for a while to add an option for that in SI2, but alas time…
Thanks ThomThom, your SI2 is great great anyway!.. I din’t know there were different acceptations of an hypothetic “True Solid” definition. But now, I will purge superfluous edges that could create overlapping faces to be accepted by SU.
I found that CleanUp3 is good tool for that…
Thanks to you,DaveR and TIG for the help.
But I still have colateral problems regarding Solid groups “gluing” . I’ll create a new post for that as it is possibly a methodology question
There is an edge here that connects to four faces - however, this is still 3d printable. Solid Inspector was made mainly with the focus of 3d printing, hence SI2 having a different way to analyse the mesh.
That being said, SketchUp’s own solid tools require solids to be “all-edges-have-two-faces” - which makes sense from a programmatically point of view because then the mesh is more predictable. This is where it would be good to have a mode switch in Solid Inspector 2.
Solid Inspector 2 will try to take face orientation into account. SketchUp is generally robust against reversed faces, but other file formats or software might not be. Even SU some times require the face orientation to be consistent for things to work properly.
So SI2 might complain about reversed faces while SU’s solid tools don’t really care about that.
But regarding faces orientation, I faced situations (for ex. after ‘exploding’ a group to glue it onto another group) were SI2 required a fix to reverse faces, but the fix ended up at unconsistant faces orientation !???
For exemple: try to explode ‘revet 3D AV G’ onto ‘AV 3D extrudé’ with: gluing rev G to AV extruded.skp (1.8 MB)
… or try the other way: exploding ‘AV 3D extrudé’ onto ‘revet 3D AV G’…
I don’t understant what’s going on ???
Hmm… that is strange one indeed. If it’d showed that there were internal faces I could understand it getting confused. But you don’t even get warnings about reversed faces while there are internal faces. I’d have to inspect the model closer to figure out what’s going on in this case.
Having seen that model from an earlier thread, it has very, very tiny errors in how the surfaces connect, much smaller than the z-fighting or coplanarity tolerances.
