Strange slicing behaviour in branched solids

[Wingcommander whpthomas]

I am wondering if anyone has run into this before - I am getting strange slicing behaviour whenever I use branched solids. Skeinforge complains about dangling overhangs, and analysing these with skeinlayer reveals layers with strange diagonal areas where there are gaps in the infill and fill areas outside the model. I tried Netfabb, and it has a pink fit and actually mades things worse - whole layers disappearing and not slicing.

I actually found the fill->start form lower left corner setting cleaned things up enough that I could get a decent print despite the errors.

Any thoughts? Maybe a module in skeinforge I should turn off?

[Eighty]

Wish I had an answer for you, but I've had some troubles with slicing certain items in Skeinforge as well.  One item, for example, looks like an inverted funnel.  SF50 placed a single layer (about halfway up) that was completely solid.  I checked the model several times, and there's nothing to suggest why it would be added.  Finally gave up and used MakerWare (which sliced it just fine).

 

I've also run across strange artifacts in the slices, as well as slightly out-of-boundary runs.  Can't explain it, except to say that every slicer will have it's quirks.

 

Have you tried rotating the model a bit?  I've had limited success with that, but it has worked a few times for me. 

[Wingcommander whpthomas]

Have you tried rotating the model a bit?  I've had limited success with that, but it has worked a few times for me.

Yes I triad diagonal layouts, but no success. This is the second time I have run into this, last time it was the rear forks - once again at the junction of a branched solid.

I just played around with different fill settings until one came out with the least amount of defects, and I printed that.

[Jetguy]

> I tried Netfabb, and it has a pink fit and
> actually mades things worse - whole layers disappearing and not slicing.

That would seem to indicate a problem in the STL right? I know you
posted pics but how about the STL? Just saying let someone else check
the STL before we go nuts trying to determine why it sliced OK once
and won't now? Sorry that seems to be what you are saying with the
picture of a good printed one and then a bad one?

On Jan 19, 3:32 pm, Wingcommander whpthomas <m...@henri.net> wrote:
> I am wondering if anyone has run into this before - I am getting strange
> slicing behaviour whenever I use branched solids. Skeinforge complains
> about dangling overhangs, and analysing these with skeinlayer reveals
> layers with strange diagonal areas where there are gaps in the infill and
> fill areas outside the model. I tried Netfabb, and it has a pink fit and
> actually mades things worse - whole layers disappearing and not slicing.
>

> I actually found the* fill->start form lower left corner *setting cleaned

> things up enough that I could get a decent print despite the errors.
>
> Any thoughts? Maybe a module in skeinforge I should turn off?
>

>  branched-solid.jpg
> 320KViewDownload

[Eighty]

Speaking of Netfabb, I've had a lingering question that maybe someone could answer.

 

I always use solid modeling for the parts I print.  So I've assumed, perhaps falsely, that there should never be issues with non-manifold items.  I use AutoCad, and I believe it runs a check during the export to STL.

 

My question is:  If you use solid modeling, you always have manifold parts - right? 

[Wingcommander whpthomas]

No they both had problems slicing, the top one has just been cleaned up. It probably is an STL problem, but it has only happened with branched solids - i.e. where two pipes become one. I can do these a different way using skinning in future which doesn't appear to cause this problem - branch solids are just quicker.

[Wingcommander whpthomas]

I always use solid modeling for the parts I print.  So I've assumed, perhaps falsely, that there should never be issues with non-manifold items.  I use AutoCad, and I believe it runs a check during the export to STL.

 

My question is:  If you use solid modeling, you always have manifold parts - right? 

Yes I was expecting the same, and this is the first feature in my CAD program I have used that has created errors like this. The program uses the ACIS modelling engine underneath, so its pretty heavyweight in that regard. However I suspect that it has something wrong with the branched solid feature, because sometimes when I add constant curve champers in the vicinity of this feature I get manifold errors.

[Jetguy]

Did you use netfabb online or local. I also assumed unlicensed (trial)
version?

I have Netfabb engine for Ultimaker which is Netfabb basic so I have
more features for repair than the demo allows.

Trying now and I'll carve one and see what happens. Fingers crossed.

[Wingcommander whpthomas]

Did you use netfabb online or local. I also assumed unlicensed (trial)
version?

I just downloaded the trial one - haven't run into any problems before this that warranted using it - so I didn't want to spring for the cash if I didn't need to.

[Eighty]

You never know..maybe Skeinforge thinks your design is a bit weak there, and decided to add a gusset for free!

[Jetguy]

Dumb question I know, but how do you attach a file via the web
interface or is that email only? Sorry for being forum ignorant.

[Wingcommander whpthomas]

Dumb question I know, but how do you attach a file via the web
interface or is that email only? Sorry for being forum ignorant.

I attach files from the web interface,  with the "Attach a file" link

[Jetguy]

Sorry, I was using the old groups view and an older browser. I know, my fault.

The main trick was "close all holes" a feature not present in the trial I think. Let me know how this works and then you can email whatever other ones you need fixed.

[Eighty]

For some odd reason, my iPad web browser doesn't have any file attachment options in this forum.  I have to get on a laptop to post with an attachment.  Odd...

[Jetguy]

I found I like the old group view which seriously limits the features
but IMO is easier to use. Thus why I never saw the buttons, it's not
in that view.

[Andy]

The Nettfab cloud system has more functionality and better repair logic than Nettfab basic.  Wingcommander, give it a try in the future when this happens.

I'd run your STL though it now, to see if it can fix it, but I'm typing this from my phone so I can't. 

-Andy

--
 
 

<R4 Forks (repaired).stl>

[Sean J]

I second Andy's idea. I use the NetFabb cloud service all the time because I don't have a solids modeling app.

~Sean

[Wingcommander whpthomas]

I second Andy's idea. I use the NetFabb cloud service all the time because I don't have a solids modeling app.

So I uploaded it to Netfabb cloud and sliced the fixed STL. I got streams and streams of error messages, but the actual slices in skein layer actually look quite good.

Warning, the triangle mesh slice intersects itself in getLoopsFromCorrectMesh in triangle_mesh.

Something will still be printed, but there is no guarantee that it will be the correct shape.

Once the gcode is saved, you should check over the layer with a z of: xxx.xxx

So whatever these errors were, they haven't really effected the slicing like the previous errors did - so Nettfabb cloud is definitely better than the trial version.

[Jetguy]

In the file I uploaded, I specifically removed "self intersections"
but the big fix seemed to be "close all holes".
I personally really like(d) Netfabb engine but was using it with an
Ultimaker. I think the fills are amazing as it's a spiral infill. They
were supposed to Make one for Makerbot have haven't seen a release.
It has huge potential, and once they got it controlling the bot,
pretty impressive. Basically an equal alternative to Rep-G in my mind.
And even just haing it around right now to fix STLs standalone for
those times when your in an internet connection free area if you know
what I mean.

There are parts of that application that were light years ahead of
anything I've seen, just wish they had worked out the compatibility.

On Jan 21, 4:14 am, Wingcommander whpthomas <m...@henri.net> wrote:
> > I second Andy's idea. I use the NetFabb cloud service all the time because
> > I don't have a solids modeling app.
>
> So I uploaded it to Netfabb cloud and sliced the fixed STL. I got streams
> and streams of error messages, but the actual slices in skein layer
> actually look quite good.
>

> *Warning, the triangle mesh slice intersects itself in
> getLoopsFromCorrectMesh in triangle_mesh.*
> *Something will still be printed, but there is no guarantee that it will be
> the correct shape.*
> *Once the gcode is saved, you should check over the layer with a z of:
> xxx.xxx*
> *
> *