1 (edited by grogbro 2013-04-30 14:02:57)

Topic: the BOX

Does anyone know why my buddies solidoodle wont print, or 'slice'? this simple box test I gave him. I made it in Maya and imported it into zbrush. Exported it out of zbrush as an STL.

Grr, ok, I cant post a link or a file. When are you allowed to post a link?

Post's attachments

6sided.JPG
6sided.JPG 61.2 kb, file has never been downloaded. 

You don't have the permssions to download the attachments of this post.

2

Re: the BOX

upload it to cloud.netfabb.com and try slicing the resulting file.  I'm guessing it's not manifold, meaning there are holes in the model somewhere.  Do a search here for the term manifold and you'll find more information.  I don't use Maya or Zbrush so I can't help in that respect. 

5 posts and you can post links.  Spam prevention technique

3

Re: the BOX

thanks for the reply! I just uploaded it to the netfabb cloud and it says it uploaded successfully and is "Manifold and oriented"

View: sixsidedtest_fixed.stl
Filesize: 7800.08kB
Triangles: 159744
Number of shells:1
Surface area: 4.39 inch2
Volume: 0.42 inch3
Outbox: 0.8 x 0.8 x 0.8 inch
Mesh is manifold and oriented.

Not sure what the dealio is

4

Re: the BOX

Is it possible that Netfabb 'fixed' the model? It didnt tell me that it did anything to it. Would it give me a message somewhere telling me that it corrected some error in the model?

Thanks again!

5

Re: the BOX

weird, both files that he has tried failed so I am not sure if its an issue with my zbrush process or the software. netfabb takes in the files fine. So I wonder if its fixing them. Although it doesnt tell me.

(working up to that 5 posts! Almost there!)

6

Re: the BOX

Netfabb doesn't really tell you anything, just that the file is done and go download it.  Are you saying that sixsidedtest_fixed.stl won't slice either?

7

Re: the BOX

What's the error message? Slic3r or skeinforge?

8

Re: the BOX

Netfabb result page will have two boxes on the right, top box will be your original and bottom will be the fixed that you download.  If you look at the details there you'll see what it's done and give you a hint as to what was wrong with yours if anything.

9

Re: the BOX

yay! I can post a link!

So as far as I can tell it didnt have to fix anything?

10 (edited by grogbro 2013-04-30 16:34:48)

Re: the BOX

http://cloud.netfabb.com/?key=7d721fa47 … ixedfile=1

Is anyone else able to take this into their solidoodle software and tell me if it's printable?

My friends printer gets this error on the Box.

"NameError: Global 3 vector"

11

Re: the BOX

sliced fine in my skeinforge.

12

Re: the BOX

sliced fine in Slic3r .99.  A bit unwieldy with all the tri-s, but not an issue it seems.

13

Re: the BOX

thanks for trying it!  Do you think it needs to be lower poly for the solidoodle to work with it?  I've seen some pretty high res examples get printed with it, so I dont think thats the issue. (?)

14

Re: the BOX

The error you're getting is a python error, but I don't know enough about it to troubleshoot.  Why don't you post the computer O/S, specs, and what software you're trying to slice with including version and maybe one of the software guys can help out.

15

Re: the BOX

There's nothing wrong with your model, aside from the unsupported overhangs you'd get on the sunken in areas would look bad.  That's not why you're getting your error.

16 (edited by cckens 2013-04-30 16:44:38)

Re: the BOX

Original mesh shows over 15k holes in MeshLab.  When you unify duplicated vertices, it closes them all up.  There are a couple of self-intersecting tri-s in one of the corners, but netfabb closed them up in its fix.  Either way, your issue is the mass number of triagles that make this up.  You could probably decimate the mesh a little and that would shrink the file size and the number of tri-s that conflict.

EDIT:
Lowering the number of triangles would simplify the mesh.  Depending on the size of the final output, you could probably do quite a bit as the fillets don't quite need to be as high a resolution to print accurately... unless you're going for a 8x8x8 box.  Still decimation can't hurt too much, especially for something coming out of zBrush or Maya

17

Re: the BOX

I gotcha, I will try to decimate it more and reupload it. Thanks!

18 (edited by grogbro 2013-04-30 16:55:49)

Re: the BOX

I decimated it by 40% and it really took it down.

http://cloud.netfabb.com/?key=cf1f67dea … ixedfile=1

I dont trust decimation master much in zbrush as it gives weird results where it tries to 'chunk' down the model. For instance in one corner of the decimated mesh there are just thousands of triangles bunched up together. the rest of the decimated mesh is fairly low poly.

19 (edited by cckens 2013-04-30 16:56:47)

Re: the BOX

Download MeshLab while you're at it (meshlab.sourceforge.net).  It's a neat little tool that can help to analyze where you may be having issues in modelling.

You'll want to at least use the Fill Hole feature (Edit->Fill Hole), the Self Intersecting Facets Filter (Filters->Selection->Select Self Intersecting Facets) and the Manifold Edge Filter (Filters->Selection->Select non Manifold Edges).  I use these to view the mesh and see where I may have modelling issues.  It makes it easier to know where I'm having issues modelling.

I use Carrara myself, so while not CAD, it still will output some good stuff.

EDIT:
BTW... MeshLab will take an OBJ file and translate it to an STL, so you wouldn't necessarily have to use the Maya->zBrush pipeline, unless you want to add detail to the model.

At 3x scale, this model took 21 seconds to slice as opposed to the 2 minutes for the other mesh, and looks just as clean.

20 (edited by grogbro 2013-04-30 16:58:18)

Re: the BOX

Nice, I will check that out tonight. wink thanks for all the help!

look at the icky stuff decimation in zbrush did to the model...bleh!

Post's attachments

ick.JPG
ick.JPG 60.73 kb, file has never been downloaded. 

You don't have the permssions to download the attachments of this post.

21

Re: the BOX

Saw that in MeshLab, but it doesn't seem to bother the slicers much.  Good luck with your modelling!

22 (edited by grogbro 2013-04-30 17:03:19)

Re: the BOX

Here is my spider model on shapeways that they printed for me, and it worked out more than fine.

http://www.shapeways.com/model/757615/b … =my-models

for the solidoodle does it matter its oriented on its butt? or does it need to be on its legs, legs down on the ground

23

Re: the BOX

You can re-orient in RepetierHost (which is the printing software I use) so it doesn't really matter, but you may have to enable support for the model as the legs might be problematic.

24 (edited by grogbro 2013-04-30 17:27:31)

Re: the BOX

during these checks in MeshLab...

" the Self Intersecting Facets Filter (Filters->Selection->Select Self Intersecting Facets) and the Manifold Edge Filter (Filters->Selection->Select non Manifold Edges)."

...if there are highlighted red areas, when you hit Apply, does it fix them, or is there a fix button in MeshLab I need to hit to fix "seld intersecting facets" ?

For reorienting the file I was using netfabb and rotating it on its X axis, that seemed to work.  So I think my process will be  Zbrush -> stl -> netfabb to reorient -> meshlab to fix errors -> netfabb cloud to doublecheck -> get final from cloud. Does that sound like a good pipeline?

25

Re: the BOX

Self intersecting facets are a bit less of an issue than you'd think.  For the most part, if there are self intersecting facets that don't break the working plane of the facets, it isn't an issue as they don't deform the mesh.  Slicers look at this and just consider it mostly clean (not always, but most of the time).  You have issues with this when you have facets that cross at or near perpendicular.  The slicer doesn't always know what to do with them, so it either goes crazy or just doesn't slice it, leaving holes in the mesh. 

A single manifold object is always the best slice.  However, if you have multiple manifold objects that are close (but not intersecting) it slices OK most of the time.

The red highlights are selected facets.  They are not fixed, just selected.  You can attempt to delete the self intersecting facets (Delete key) then use the Fill Hole command, but that may have "interesting" consequences.  Best bet is to learn where your errors are occurring and then go back and check the model in facet mode.  In zBrush this may not be as easy, but then again, I don't know zBrush well.

The best pipeline is one that works for you.  Your suggestion seems a bit... long, but it would help to find and avoid mesh issues.  Like I said before, the best plan is to learn where your model is faulting, then fixing your modeling process so that you avoid those pitfalls.  Decimate will help to clean a model up, but sometimes, you need the tri-s to make a model look good.  I suggested it before because a box like that has a lot of open area and not a lot of detail.  Best export for that model would have been to take it wholesale from Maya (export as an OBJ, or STL if Maya can do that) and check that in MeshLab.  I would bet your tri-count would have been a lot lower.  For a zBrush detail model, I'd keep it whole without decimation to get the most detail out of it.

Re-orientation in netfabb works, but I think that if you're going to produce objects for printer output, I'd download either ReplicatorG or RepetierHost to install on your machine as these will allow you to re-orient in real time, and then re-save the STL in the new orientation.  Essentially, you should probably have a machine capable of 3d printing even if you don't have a printer.  This would allow you to output some pretty awesome stuff (if that spider is any indication of your modeling skills).  Look up posts by ysb as he's the resident guru on high quality figure models.