Monthly Tech-Tip | No tracking! No ads! |
Infill and support are structural elements that 3D slicer software uses (e.g. to fill the interior of solid parts or support contours over empty space)
Key phrases linking here: infill and support, printed support, print supports, print support, infill - Learn more
When solid models are 3D printed they are not typically printed solid (except for very small objects). Slicer software generates infill. To be infilled, objects must be water-tight. Infill is only visible while an object is being printed, it cannot be seen on already-printed objects. Infilled objects can be 90% lighter than solid ones. Slicer software can typically generate a dozen infill different patterns (e.g. honeycomb, rectilinear, triangles, grid). Along with the pattern the slicer expects a percentage of fill density (typically defaulting to 10-20%). Infills typically have walls of the same thickness as the nozzle, commonly 0.4mm - they can thus be quite delicate. Proper infill generation depends on the 3D printer being in good adjustment to assure that successive 0.4mm laydowns are exactly on top of the one below. Slicers can generate solid layers top, bottom and on all sides of infills.
The minimization of infill and support is a key aspect in the reduction of print times and reducing waste material. Experience is needed to learn when it is best for an object to be printed hollow, hollow with support or solid with infill. Infills generally print more trouble-free than supports. On larger objects, infills can drastically increase print time, even though fill density might be quite low. Infills of low density (e.g. grids where the lines are 5mm or more apart) can be quite practical - but there is a limit to how wide a space the top layer can span without sagging too much.
Strategic placement of flanges on parts can also really minimize print times. Breaking prints up into multiple pieces having flanges can not only enable each to print without support but also can make the parts reusable (since they are clamped together they can be disassembled). Sometimes brims can be printed against the bed (e.g. 3-4mm wide with no separation) enabling even 0.8mm thick walls to be printed right on it (the brims hold them down, otherwise they would release from the bed). That brim can then be used as a gluing surface and clamping flange (and can be removed on plaster-facing surfaces when the glue is dry).
These are quarters of a block mold piece for a Medalta Potteries ball pitcher. The whole piece was far too big to print so I had to break it up into quarters. Quarter #1 (top right and bottom left) was made by shelling (hollowing) the whole thing in the CAD drawing process and quartering in the slicer. The other piece (top left and bottom right) was made by quartering and shelling each individually in the CAD software (Fusion 360). Quarter #1 has a dangling corner (front left of top right) so I had to print as shown so that infill would support it. Quarter #2 mates with a wall and that supports the whole curved mating edge (enabling upright printing). Notice also that the matting surface is not planar on quarter #1 (top right). And its inside surface as print artifacts. Quarter #2 was printed with the back-side down, thus only printed support was needed in the narrow channel where the rubber wall will pour. Surface quality is much better and it printed in 9 hours instead of 14 (these pieces are quite large).
This is the bottom section of a Medalta ball pitcher model, it is part of a project to create an all-in-one case mold. It has been printed upside down and printed support has been generated by the slicer software (any contours having an angle of less than 40 degrees are supported). The printed shell is light, it weighs only 47g (the support weighs 78g). The outer rounded surface is the finished one, in this orientation maximum quality has been produced. Had this been printed right side up support would have been generated all the way around the outside. And, because the finished surface would have been created on top of support, its quality would have been poor. These supports are also excellent opportunities for recycling PLA filament.
In a 3D print, a small amount of infill can greatly outweigh the generation of a large amount of printed support. This can greatly reduce print time and filament consumption (this is printed with 0.8mm thick walls).
Far left: The green part is the support my slicer has generated in order to hold up the flat top flange. This is not only a waste but problems are likely with adhesion on the lower support sections where tiny points are printed on the steeply included surface to start each.
Middle: The same piece has been edited to terminate at the top with a triangular cross section. This is a 40 degree angle (as can be seen on the far right) - the slicer has been configured with an overhang threshold of 39 degrees, that means it does not need to generate support. It does infill that section, but the increase in time is minimal.
I am 3D printing half of the bottom section of the outside shell assembly of an all-in-one case mold (to make a plaster mold for casting pottery). There are multiple factors to take into account to make this print quickly with minimum material and yet be strong.
Left are case molds, they are made by 3D printing the positive profile on a backplate (with holes for natches). These are secured into slotted rails. Right is a block mold, it is made by 3D printing the profile of a working mold with integrated rails. This one is printed vertically in four pieces. It is held together and straight with printed brackets. We pour rubber into these to make case molds. Each method has advantages and issues.
-Case: Faster to print. Easier to draw. Joins cast as easily removed bumps on the working molds. This is only suitable for prototyping, making one working mold.
-Block: Much more attention is needed in printing, there are more issues with orientation of print, infill, printed support, multi-piecing, fit and seam-filling. 3D drawing of these is more difficult. And block molds are bigger because they are molds of molds. They also need to be more precise to merit the cost of the rubber.
This is a Medalta Potteries medium-sized ball pitcher block mold, version 2.0 - it has a more oval body shape. Upper left is the top section (actually, it is half of the top section, a base will fit on to create a three-piece working mold. There is to 3D print something this large in one piece in a consumer 3D printer. Even if it did it would require 50 hours of print time! Also, working surface quality is affected by the orientation of printing (especially if a surface is down-facing and on top of printed support). Further, a large print would almost certainly warp and corner-lift during printing. Cutting it into four pieces and hollowing them individually (in the CAD software, lower right) solved all the problems. Each of the pieces is still quite large, taking 10+ hours to print. But they can each be hollowed individually and rotated to the optimum position for the best finished surface (and print speed). It is amazing how well these four pieces fitted together. This approach really paid off because I made a mistake - and I only needed to reprint one of the pieces!
Top: The left one has a rectilinear fill, the right a square grid fill. Notice the fill on the left has buckled part way up - this piece is too tall for that type. While the grid fill on the right is far more stable, it is not as easily removed (although it can stay in place here).
Middle: The piece on the right printed in half the time because the only fill needed is at the bottom. The fill is rectilinear and easily removed - it affects the smoothness of the surfaces but they are not a finished ones so it does not matter. However that method is risky, notice the failed print on the lower right - upward pull of the flat section pulled it away. This happened twice more so I chalked up the one success to luck.
Bottom: Took about 16 hours. Like others, the grid of printed support out past the edge - that part has to be cut away with a sharp blade knife to enable mating with the other pieces.
Glossary |
0.8mm thickness
This is an ideal wall thickness to print on consumer 3D printer since to normally creates sufficient strength and prints in one two passes |
---|
By Tony Hansen Follow me on |
Buy me a coffee and we can talk