FDM printing with 3d layers

[albert 2]

I thought I would bring this into the CURA thread, because I think it is a completely new idea. At least I have not seen this before.

The basic idea is to slice the object in such a way that the layers follow a 3d path. The z axis moves up and down during the printing of a layer. This way the fibers of the material follow the form. Might be a new way to get around overhangs and also might make parts that are much stronger. One could also imagine doing thicker layers without the ugly stairstepping effect on top layers with a bent surface. The results look good, especially for the crappy printer he uses in the video:)... Dont know if this can be done with an unmodified ultimaker mechanism.

At least the head would have to be modified with more clearance and a longer nozzle. I also like the idea of doing a support structure, stopping the printer, putting in a release agent, and then continuing on the support layer.

Food for thought!

 

www.topolabs.com

[Dim3nsioneer 557]

There are a few posts about Topolabs already here in the forum (in case you haven't seen them):

http://umforum.ultimaker.com/index.php?/topic/5252-topolabs-nonplanar-fdm-fff-toolpaths-for-improved-desktop-3d-printing-quality/

http://umforum.ultimaker.com/index.php?/topic/5203-training-required/&do=findComment&comment=46155

I think it's a really nice idea but it makes a quite pointed nozzle necessary. A standard UM nozzle can be used as well but will have its limitation with steeper slopes.

One has to keep in mind that o print an object this way does not make it necessarily stronger. E.g. one of the weakest interfaces of a 3D print is the one between outer shell and infill (even if you have a good overlap). If you manage to make the slicer engine defining layers where these interfaces are slightly shifted from layer to layer, you can increase strength significantly.

But I'm pretty sure some 3D printer manufacturers might go into those directions. I'm looking forward to what Daid is going to do with such ideas in Cura's future...

 

[aviphysics 9]

It's a cool idea and I am paying attention to where this, but I also feel like the part they printed the conventional way is not representative of a good quality print.

 

[Daid 306]

I'm getting the feeling they are not slicing, but designing the toolpaths straight from their application. Fine for their goal, but a generic solution.

Automated slicing in curved 3D layers would be pretty damn complex.

 

[Dim3nsioneer 557]

Automated slicing in curved 3D layers would be pretty damn complex.

 

Maybe it's easier to go the other way round? Applying (the inverse of) a function to the model (model transformation), slice it normally and applying the same function to the toolpath?

 

[aviphysics 9]

Maybe it's easier to go the other way round? Applying (the inverse of) a function to the model (model transformation), slice it normally and applying the same function to the toolpath?

 

Do you mean something like designing a generic flat pattern and then applying a transformation to warp the pattern up or down in the Z direction, in order to follow the object surface?

 

[gr5 2,167]

Sort of. He means transform the part flat, then slice, then transform it back to the original shape.

This brings to mind Joris's only "truly" 3D printed cup:

http://www.thingiverse.com/thing:75735

which was done as a cura plugin that transforms the Z positions. Pay attention to the colors! Think about how you would do those colors on a UM with only one nozzle!

 

[gr5 2,167]

The part printed in that video could have just as easily have been printed flat, then heated over a form and pressed into any shape. Still the concept is very cool. This wedding dress was printed on a UM original in flat sections and then shaped I assume with heat:

 

[Dim3nsioneer 557]

gr5, you caught my thought quite well. However, the link you provided doesn't seem to fit the topic? :???:

 

[gr5 2,167]

r, the link you provided doesn't seem to fit the topic

 

Fixed!

 

[Dim3nsioneer 557]

I don't understand. Is she too flat? Not sufficiently 3D? It was printed with an Ultimaker and then heat was used to shape over a form/mold.

 

Uhhmmm... the PICTURE...

 

The video above prints directly on the form, but why not print first flat, then mold with heat afterwards?

 

I guess it depends on the amount of deformation and if you require a certain thickness of the print after deformation.

 

[Daid 306]

Maybe it's easier to go the other way round? Applying (the inverse of) a function to the model (model transformation), slice it normally and applying the same function to the toolpath?

Appying a function to straight polygons which results in curved polygons totally not falls in the "easy" category ;-)

 

[Dim3nsioneer 557]

Appying a function to straight polygons which results in curved polygons totally not falls in the "easy" category :wink:

I never used the word "easy", I wrote "easier" which means something like "less difficult" without actually stating how much less difficult... :lol:

No, seriously, I see your point... let me just think aloud... if the polygons (triangles in an STL, right? I'm absolutely no expert for the stl format) are sufficiently small, let's say much smaller than the deformations applied by the transformation function, straight polygons would be deformed in size but would stay straight in first approximation. Larger polygons however would have to be split up in sufficiently small polygons before transformation. There should be some good algorithms for this step used in e.g. FEM.

If you think I'm writing bs, just let me know please. I can stand it... :smile: