Adding mirrored object in makerware?

[c f]

I've come across an odd problem. Makerware (currently running 2.2.2.57 on ubuntu) doesn't have an option to mirror objects. To work around this, I load the object into ReplicatorG 0040, mirror the STL, and re-save it as a different STL file. I then add it as a second object using makerware (I am trying to print both the original and the mirrored object at the same time). For some reason, when I click 'make', it only slices the original object, ignoring the 2nd, mirrored object. Has anyone come across something similar? I can actually save the STL file with both objects using Makerware, and then slice it using ReplicatorG, and it works fine. Definitely an odd bug.

-Chris

[Damian Gto]

Do NOT use RepG to do this. You will end up with a useless object that is broken.

So its not a bug in makerware. you can only print meshes that is good.

Use an other tool for this, like netfabb, meshlab etc.

[c f]

So the issue really is that ReplicatorG is corrupting the STL file when it mirrors it, and then makerware is just ignoring the broken file?

--
You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.
To unsubscribe from this group and stop receiving emails from it, send an email to makerbot+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Joseph Chiu]

Possibly.  I know that with TurboCAD, some STL's that I modified and then saved inside RepG would sometimes not open anymore in TC. I haven't investigated this in detail, though.

[Dan Newman]

On 12 Aug 2013 , at 9:05 AM, c f wrote:

> So the issue really is that ReplicatorG is corrupting the STL file when it
> mirrors it, and then makerware is just ignoring the broken file?

When you mirror the object in RepG by scaling by -1, you end up with the
surface normals all inverted. That's not broken and various slicers handle
it just fine. MakerWare, on the other hand, has some conventions
regarding STL surface normals which this runs afoul of. Neither RepG nor
MakerWare is "broken" -- they're just different. MakerWare's convention
here may be inconvenient but not necessarily wrong.

Dan

[Clinton Hoines]

I was just going to past exactly what Dan did. Now if you have inverted normals just load your file into Meshlabs and under the filters drop down menu you find another drop down for normals and invert them easy.
Or you can mirror your part in Netfabb and it mirrors properly in the first place no inverting needed. ;)

[Dan Newman]

On 12 Aug 2013 , at 11:02 AM, Clinton Hoines wrote:

> I was just going to past exactly what Dan did. Now if you have inverted
> normals just load your file into Meshlabs and under the filters drop down
> menu you find another drop down for normals and invert them easy.
> Or you can mirror your part in Netfabb and it mirrors properly in the first
> place no inverting needed. ;)

And a slight quibble: the issue isn't that RepG mirrors correctly or incorrectly.
Rather, RepG doesn't mirror at *all*. What folks are doing is using RepG's
scale function to get an effect very similar to mirroring but not quite. "Not
quite" in that scaling by -1 means inverting the normals as well. That's
correct behavior when applying a scaling transform in matrix math land. But
that's not a mirror transform. But then RepG doesn't claim to be mirroring --
it's scaling and tells you that it is since the button is labelled "Scale".
So the improper behavior here is actually human behavior: using a scaling
function for mirroring. It'll work with some slicers but not with others.

Dan

[Damian Gto]

Yes it does.

Some bad STL makerware can read, but not when they got that broken.

Even if you run them through a "fix" program you can not be sure its okey.

RepG is NOT made to make STL good files. IF someone tell different , they do not know how it works.

You need a lot more advanced program that is MADE for 3D object and make a TRUE mirror of the object.

I have seen to many times people mess with 3D object and have no clue what they are doing.

The sad fact is that few know how to make a good 3D object that they put on Thiniverse. If I download any object from that place I make always use a program to check the object so its okey. I can tell you most of the time they are bad for 3D printing. You can also see there is mirror object that people done and most of the time they are bad.

This is why I often tell people that they should learn to make there own objects. It will save time and plastic :-)

Also if you do you have fulll controll over them and its very easy to mirror them if needed. :-)

[Richard]

My guess is that when you are mirroring it (make a reverse?) The second file is turned inside out from the first and is no long manifold/watertight.  If that is the case, it will not slice.  Send the mirrored stl to cloud.netfabb for fixing.

When it is turned inside out, it will look fine on most viewers but every triangle will have been flipped so the "inside" face is now on the outside of the object and it has infinite volume.  On netfabb desktop, inverted triangles show up as Maroon, good ones are grey or green if the object is selected.

[Dan Newman]

On 13 Aug 2013 , at 8:19 AM, Richard wrote:

> My guess is that when you are mirroring it (make a reverse?) The second
> file is turned inside out from the first and is no long
> manifold/watertight. If that is the case, it will not slice. Send the
> mirrored stl to cloud.netfabb for fixing.

In general, that is what happens when you invert all the surface normals: the inside
is now interpreted as the outside and vice versa. Some slicers pay attention to that
and others do not. I suspect that those in the latter camp are just then looking
for consistent orientation of each poly forming the surface. Skeinforge is an example of a slicer
that falls into the latter camp which is why people encounter in internet searches
advice to use RepG's scaling op as a means of mirroring. It works within the confines
of RepG + Skeinforge but at the end of the day doesn't produce truly a mirrored piece.

As far as a STL tool, RepG is generally okay for actual scaling and moving about
the origin but that's it. And the STL files it outputs are always in human-readable form
and not the more compact binary form. Thus the STLs RepG writes back out take up
considerably more disk space. Not a problem unless you will be uploading it or
e-mailing it. Net, Net, people who need to do even simple manipulations on STL files
should invest a little time in learning to use a more capable tool (free version of
Netfabb, Blender, or whatever). Nothing new with that advice -- has been said plenty
in this forum over the past couple of years.

Dan