Infill questions

[axsdenied 0]

Few questions regarding infill:

1. What is the best way to do 100% infill?

For example, I can increase the number of walls (and hence have shorter/less "diagonal lines") or I can reduce the number of walls to 1 and use "diagonal lines" to do the filling?

Does it make a difference regarding the final strength?

2. Does the shape of infill have effect on the strength? For example slic3r offers lines, concentric, honeycomb etc.

3. Is there a "magic number" of infill that gives higher strength than similar, slightly higher values?

For example, at 25% infill Cura starts laying infill lines in one direction which changes every layer. Does this change have a significant effect on the strength, say compared to 24%?

 

[nallath 677]

The answer to pretty much all questions is; We don't know.

We've not tested this, mostly because it's very hard to even quantify 'strength' (and what makes one object strong might dramatically fail for the next one).

 

[axsdenied 0]

Thank you.

Somehow I expected that answer. The mysteries of 3D printing continue :smile:

 

[aviphysics 9]

I have been beginning to wonder how much strength 100% infill actually offers over a spars infill, like 10%.

From a mechanics perspective, it probably increases the objects resistance to pulling or compressing, but not for bending.

For bending, most of the load is carried by the outer walls, instead of the core. The main function of the infill is to help the skin hold its shape, so it can't easily collapse or expand.

For stretching and compressing, the load is more evenly distributed. The most important thing is the total cross section of material that need to be compressed or pulled apart.

In practice, I am beginning to feel like anything over 20% infill is generally a waste.Unless there are special circumstanes, it is best to just design a part with slighty large diameter or increase wall thickness a little.

As for infill patterns, I would like to see some alternatives to the grid we have now. The grid leaves week spots and I don't like the look it lends to translucent material. One thing I would like to try is just randomly rotating the grid at each level, making a sort of rat's nest of overlapping fibers that fills the object by some %. Some other more complicated ideas I have scene on these forums are a pattern that would just form ribs along the walls or an algorithm that would do the grid but bend the lines near the wall, in order to form a curve that is tangent to the wall at the point of contact.