3d printing hacking: tailoring strength of infill material

Talk about anything.
Post Reply
outsider
Posts: 49
Joined: Tue Mar 12, 2013 2:08 pm

3d printing hacking: tailoring strength of infill material

Post by outsider » Sat Sep 14, 2013 10:17 pm

In an effort to try and control the internal fill material of a part (both direction and amount) I've been experimenting with some ideas and finally have something.
(This hack applies to any 3d printer in fact, not just the UP, but I'm using an Up since it's what I have)

If the generic cross pattern is either not strong enough for you, or you want to tailor the direction of strength a little better what you must do is model some features in your part that the slicing software can use to put material where you want it.

The simplest method I've found is to model internal "cuts" or "slits" in the part. For the UP printer, the slicer software will do the 2 walls around the opening, but if the cut or opening is modeled narrower then 0.1mm, the opening never really materializes (is not present in the internal structure of the print) since the material from both sides will fuse together when being printed.

So if your complex geometry part needed to be stronger in one direction then in another, you could model some "slits" in the direction you want more strength* and less (or no slits in other directions)
* More strength in comparison to the generic fill of the slicer software.

This method can not only generate strength in desired directions, you could use it to just create thicker walls (more around the perimeter, more walls)
Another thing this method can create (which is not currently possible with the UP software) is a solidly filled object.
Untitled.jpg
Original CAD file
Untitled.jpg (102.3 KiB) Viewed 12157 times
IMG_20130914_175043B.jpg
Printed parts
IMG_20130914_175043B.jpg (203.61 KiB) Viewed 12157 times
On the left is a generic filled part and on the right is a nearly solid part due to the slits modeled in the part.

One other thing to note: if the "slits" travel all the way from wall to wall, you risk showing the pattern on the outside of the part. If that's not desired, you could make sure the modeled slits begin 0.5mm beneath the surface of the part. That way all the slits will be internal to the part.

This is just the beginning of my testing with this idea. I think it has a lot of potential, and I will post more as I get more designs made.
adrian.
My Blog

User avatar
JuliaDee
Posts: 959
Joined: Sun May 06, 2012 10:00 pm
Location: Connecticut
Contact:

Re: 3d printing hacking: tailoring strength of infill materi

Post by JuliaDee » Sun Sep 15, 2013 1:08 am

Do you really think the part on the right is stronger than the one on the left? I mean, they don't fill these in... :)
bridges2.gif
bridges2.gif (13.94 KiB) Viewed 12146 times

roller
Posts: 1622
Joined: Sat Oct 06, 2012 9:09 pm
Location: Brisbane, Australia

Re: 3d printing hacking: tailoring strength of infill materi

Post by roller » Sun Sep 15, 2013 1:29 am

This is an old reprap trick to get solid fill in certain parts of your model... pretty nice work to figure it out on your own too. There are two caveats to be aware of when you are trying to trick the printer into a true solid full (feature request!!!):

-as Julie points out the hatching like a truss can be quite strong and more efficient but that also that you can end up with fissures between your infills where the slicer decides they are too tight to bother will infill - in doing so you end up with a weaker model in one direction. The magic point for infill fissures changes with layer height so do forget to experiment in the Z axis as well to find the optimum.

- This technique when done right (without fissures) has a tendency to generate overfill so be wary it can effect perimeter and top surface print quality sometimes.

Also, this technique is not just a great way to increase model strength it is also useful to control the balance of a model by being able to add mass to parts with greater density. It would be great if you could predict the changes in density (extra added mass) your method creates.

outsider
Posts: 49
Joined: Tue Mar 12, 2013 2:08 pm

Re: 3d printing hacking: tailoring strength of infill materi

Post by outsider » Sun Sep 15, 2013 2:36 am

Julia,
Excellent observation.
The strength to weight ratio is very high when you triangulate the area. This however only works when the applied load is in the plane of the triangulation.
When a load is applied perpendicular to the that, the triangulation has absolutely no effect.
Also to note is that the infill pattern that we get on the Up printers is NOT a triangulation effect! You get squares, which are not as effective as triangles in resisting the "bridge" loads you pointed to.
To go into more detail, when you fully divide a surface in triangles, you end up with only tension and compression in each arm of the triangle. Once you have squares involved, each arm of the square is in bending. Tension/compression is much easier to resist then bending.

Back to 3D printing; if the part is oriented at 45deg on the print bed, the pattern you get is a bunch of squares lined up with the edges of your printed part.
Not ideal if YOU want to control the direction of the plastic fibers on the inside of the part.

One other issue I have with the generic infill is that it is not continuous in the Z direction. If you watch the printing, one layer gets +45deg lines, and the next layer gets -45deg and it continues to alternate like that.
So the bond between two layers in the same direction is minimal (at best) since the infill layers in one direction are deposited (0.15mm or 0.3mm or whatever your layer height happens to be) with a gap between them.
It would be a different story if both the +45 and - 45 deg infill would be deposited at each layer.

With the slitting you can control the direction in the part geometry, not by relying on it's orientation on the print bed.
adrian.
My Blog

outsider
Posts: 49
Joined: Tue Mar 12, 2013 2:08 pm

Re: 3d printing hacking: tailoring strength of infill materi

Post by outsider » Sun Sep 15, 2013 2:51 am

Roller,
You bring up some excellent points.
I do agree that the fissures can lead to weakening if not done correctly. I am experimenting with a few values and so far I am getting excellent results with fissures that are 0.05mm wide (in the cad file) and spaced 1.6mm apart (so the center of one fissure is 1.6mm away from the center of the next fissure)
The Up Printer prints 2 perimeters around each fissure, and that is about 0.8mm to 0.9mm wide, so just shy of 4 perimeters between fissures seems about right.

I will take care to monitor the overfill to see if the part dimensions are affected beyond unacceptable limits.
Last edited by outsider on Sun Sep 15, 2013 3:50 am, edited 1 time in total.
adrian.
My Blog

outsider
Posts: 49
Joined: Tue Mar 12, 2013 2:08 pm

Re: 3d printing hacking: tailoring strength of infill materi

Post by outsider » Sun Sep 15, 2013 3:40 am

Did a bit more testing.
Here are some test parts showing thicker perimeter, and an internal truss.
IMG_20130914_232149.JPG
IMG_20130914_232149.JPG (68.3 KiB) Viewed 12139 times
I have measured a slight "overfill" where the perimeter grows slightly. The growth I've measures seems consistent at about +0.15mm on each side of the perimeter.
Depending on what you're making this may not be an issue. It can also be accounted for in the CAD model so the final part is the right size.

To note; once a fissure is made, there is about 1mm of material on either side of the fissure. So the truss on the right part is about 2mm wide. No way to make it narrower.

I think the best part of all this the ability to internally "truss" a part in the Z direction with solid material. Normally in the Z direction the only material that can resist breaking is the strength of the inter-laminar bonding of the plastic fibers(deposited filament). And really only the perimeter 2 layers, since I would not count on the infill to resist much when stressing the layer bonds.
Imagine printing a tall (in the z direction) skinny tower. If you try to snap it, it will break at the layer bonds.
Now I can make internal diagonal solid pillars (reinforcements) that can resist forces in the designed directions.
adrian.
My Blog

User avatar
JuliaDee
Posts: 959
Joined: Sun May 06, 2012 10:00 pm
Location: Connecticut
Contact:

Re: 3d printing hacking: tailoring strength of infill materi

Post by JuliaDee » Sun Sep 15, 2013 3:50 am

outsider wrote: Imagine printing a tall (in the z direction) skinny tower. If you try to snap it, it will break at the layer bonds.
Now I can make internal diagonal solid pillars (reinforcements) that can resist forces in the designed directions.
Hey, even I can see the usefulness of that! :-) Nice work.

I notice you mentioned a couple of times "the 2 perimeter layers". FWIW, the Up software allows you to specify up to six layers of perimeter thickness. Three is the default, but I often print with four. Even for non-structural stuff (e.g. sculptures) it sometimes seems to give a nicer part. And if I know I'm going to be drilling or tapping I always use an extra-thick perimeter.

julia

outsider
Posts: 49
Joined: Tue Mar 12, 2013 2:08 pm

Re: 3d printing hacking: tailoring strength of infill materi

Post by outsider » Sun Sep 15, 2013 3:53 am

I thought we only had control of the top/bottom layer thickness, not the side wall thickness (which I thought was only 2 layers).
I do have the "Part>Surface" set to 6 layers but as far as I knew that setting only controlled the top/bottom surface and not the side perimeter.

What setting specifically are you referring to Julia?
adrian.
My Blog

User avatar
JuliaDee
Posts: 959
Joined: Sun May 06, 2012 10:00 pm
Location: Connecticut
Contact:

Re: 3d printing hacking: tailoring strength of infill materi

Post by JuliaDee » Sun Sep 15, 2013 3:58 am

outsider wrote:I thought we only had control of the top/bottom layer thickness, not the side wall thickness (which I thought was only 2 layers).

What setting specifically are you referring to Julia?

adrian
Ah, you may be right... I always assumed that the "Surface" setting applied to all surfaces, not just "top/bottom". So how does it treat curved surfaces if it's only top/bottom?
Screen shot 2013-09-14 at 11.56.17 PM.png
Screen shot 2013-09-14 at 11.56.17 PM.png (34.61 KiB) Viewed 12136 times

outsider
Posts: 49
Joined: Tue Mar 12, 2013 2:08 pm

Re: 3d printing hacking: tailoring strength of infill materi

Post by outsider » Sun Sep 15, 2013 4:07 am

The way I think it handles sloping walls is by the "Angle" parameter in the screen-grab you posted.
If the slope angle is less the angle specified, it's treated as a top surface, hence giving it 6 layers. I think the angle is measured from the horizontal.
adrian.
My Blog

User avatar
JuliaDee
Posts: 959
Joined: Sun May 06, 2012 10:00 pm
Location: Connecticut
Contact:

Re: 3d printing hacking: tailoring strength of infill materi

Post by JuliaDee » Sun Sep 15, 2013 4:13 am

Per the documentation, the "angle" parameter determines at what angle it starts placing "dense support" (in which case it should be in the "Support" section of the settings, which I've previously mentioned to pp3dp). I've never seen any evidence externally of any transitioning from thin to thick walls on curved surfaces.

But yeah, it does say right there that the thickness setting doesn't affect "side" walls.
Screen shot 2013-09-15 at 12.11.01 AM.png
Screen shot 2013-09-15 at 12.11.01 AM.png (107.84 KiB) Viewed 12137 times

roller
Posts: 1622
Joined: Sat Oct 06, 2012 9:09 pm
Location: Brisbane, Australia

Re: 3d printing hacking: tailoring strength of infill materi

Post by roller » Sun Sep 15, 2013 11:56 am

Sorry, when I say fissure I meant unintentional gaps between sets of new perimeters you were creating ... Because extrusion width is related to layer height the way perimeters interact with adjacent perimeters of other slots depends on layer height. So what I am saying is you will need to repeat your experiments at each layer height as your ideal spacing will likely vary for each.

If there is not gap between outsides, which is what you've been demonstrating in the first example you have a nice solid section. If the perimeters are a little thinner due to a change in layer height you may end up with a small gap and worse, because the slicer thinks the gap is too small to create an infill it will leave it completely empty (a fissure) so you get no infill nor perimeter.

This is not a flaw in your technique and I am not trying to criticise it ... it's just as you develop rules of thumb be mindful you might need to test your set of rules for each layer height ... or always print at a consistent layer height. Hopefully you will find a rule of thumb that works at all height though.

Finally, if you have a temp mod you can ever so slightly reduce overfill by by lowering the temp 10-20C provided that doesn't generate jams for you.

Post Reply