Source : tested.com
For Our Users in Egypt who are interested in modeling for 3D printing, below are some very important tips for 3D modeling.
Be Tidy and Neat in Your Modeling
Taking the time to sculpt a neat, clean computer model will prevent headaches down the road. This is particularly true of polygon models where deleting an edge, face, or vertex can quickly make a model unprintable. Using boole operations (adding and subtracting part together) is often used while building models, but can lead to messy models since two pieces of geometry are being combined or subtracted from one another.
Sloppy modeling can easily occur just in the process of figuring out how to build something. I will often build a quick, rough model to work through the layout, what parts need to be made, and how to build them. I will rebuild the whole thing as a much cleaner model based on the rough version. One of the best pieces of advice I got from my modeling mentor is, ‘don’t be afraid to rebuild something’. It sounds like a drag but rebuilding a model from scratch always goes quicker than the original and it will be a cleaner model, using what was learned from the first version.
If modeling with polygons, it’s in your interest to keep the mesh in quads (each face is four-sided) and avoid “n-gons” (in modeling, any polygon that is not 4-sided). Modeling with quads makes adjusting the model much easier, whereas n-gons will kind of mess things up. In general, any modeling program will make it easy to model in quads since any primitive (cube, sphere, cone, torus, etc) created will automatically be made out of quads.
Taking the time to sculpt a neat, clean computer model will prevent headaches down the road. This is particularly true of polygon models where deleting an edge, face, or vertex can quickly make a model unprintable. Using boole operations (adding and subtracting part together) is often used while building models, but can lead to messy models since two pieces of geometry are being combined or subtracted from one another.
Sloppy modeling can easily occur just in the process of figuring out how to build something. I will often build a quick, rough model to work through the layout, what parts need to be made, and how to build them. I will rebuild the whole thing as a much cleaner model based on the rough version. One of the best pieces of advice I got from my modeling mentor is, ‘don’t be afraid to rebuild something’. It sounds like a drag but rebuilding a model from scratch always goes quicker than the original and it will be a cleaner model, using what was learned from the first version.
If modeling with polygons, it’s in your interest to keep the mesh in quads (each face is four-sided) and avoid “n-gons” (in modeling, any polygon that is not 4-sided). Modeling with quads makes adjusting the model much easier, whereas n-gons will kind of mess things up. In general, any modeling program will make it easy to model in quads since any primitive (cube, sphere, cone, torus, etc) created will automatically be made out of quads.
Due to limitations of both the printer and material used, you must pay constant attention to wall thickness, size of details and overall size of the print. If you make a wall too thin or detail too small you will most likely have some holes and breakage on the printed model.
Wall Thicknesses
Due to limitations of both the printer and material used, you must pay constant attention to wall thickness, size of details and overall size of the print. This can be tricky considering that not all modeling programs let you build in a particular unit. Some programs may list a circle as being 10 in diameter, but what is the unit of measurement? That’s the glory of digital design; it doesn’t really matter since the model can be scaled up and down almost indefinitely. But if the material you are printing in needs a minimum wall thickness of 1mm, then units become important.
In this case, I will generally decide that 1 unit = 1 cm or 1 inch, etc and model accordingly. Even having done this, you may export the STL and rather than 10 cm long your model is 100 cm long! It all depends on how your program exports, some will allow you to specify the units, some won’t. Regardless, once the STL is opened in the slicing program you should be able to scale it to the correct size.
Due to limitations of both the printer and material used, you must pay constant attention to wall thickness, size of details and overall size of the print. This can be tricky considering that not all modeling programs let you build in a particular unit. Some programs may list a circle as being 10 in diameter, but what is the unit of measurement? That’s the glory of digital design; it doesn’t really matter since the model can be scaled up and down almost indefinitely. But if the material you are printing in needs a minimum wall thickness of 1mm, then units become important.
In this case, I will generally decide that 1 unit = 1 cm or 1 inch, etc and model accordingly. Even having done this, you may export the STL and rather than 10 cm long your model is 100 cm long! It all depends on how your program exports, some will allow you to specify the units, some won’t. Regardless, once the STL is opened in the slicing program you should be able to scale it to the correct size.
Subdivide and Conquer
3D printers only print pure geometry, so you need to make sure your model is smooth enough to print well. If you are using a polygon modeler remember to turn off any rendering shaders since they fake smoothness on the screen.
If your model looks chunky on-screen you will need to add more subdivisions which means adding more polygons. How many subdivisions to add is a tough question since the answer depends on multiple factors. A sphere may look a little blocky on screen but maybe it’s only 10 mm in diameter when printed so, at that size, it will probably print just fine. It’s also possible to go to the opposite end of the spectrum and have too many subdivisions. Going crazy with subdividing will bog down your modeling program, the STL export, the mesh repair program, the slicer and maybe even the printer.
A good example of over-subdividing is making small holes on a typical home FFF printer. If you add enough subdivisions to make a perfectly smooth hole the printer won’t be able to properly translate so many small moves. It’s better to use a lower edge count which will produce a smoother printed opening.
Adding subdivisions to something like a sphere is pretty straightforward; add more divisions, sphere looks smoother. But what if you want to make a cube with rounded edges? I could subdivide the heck out of it and it will still look like a cube--this is where you need a subdivision surface, AKA “subD” modeling--not to be confused with regular subdividing. SubD modeling is kind of the polygon equivalent of modeling with NURBS since it’s good for smooth organic surfaces but doesn’t take as much computing power. Adding a subD modifier to an existing model will add subdivisions AND a smoothing algorithm, similar to the render smoothing, but in this case the actual geometry is smoothed.
3D printers only print pure geometry, so you need to make sure your model is smooth enough to print well. If you are using a polygon modeler remember to turn off any rendering shaders since they fake smoothness on the screen.
If your model looks chunky on-screen you will need to add more subdivisions which means adding more polygons. How many subdivisions to add is a tough question since the answer depends on multiple factors. A sphere may look a little blocky on screen but maybe it’s only 10 mm in diameter when printed so, at that size, it will probably print just fine. It’s also possible to go to the opposite end of the spectrum and have too many subdivisions. Going crazy with subdividing will bog down your modeling program, the STL export, the mesh repair program, the slicer and maybe even the printer.
A good example of over-subdividing is making small holes on a typical home FFF printer. If you add enough subdivisions to make a perfectly smooth hole the printer won’t be able to properly translate so many small moves. It’s better to use a lower edge count which will produce a smoother printed opening.
Adding subdivisions to something like a sphere is pretty straightforward; add more divisions, sphere looks smoother. But what if you want to make a cube with rounded edges? I could subdivide the heck out of it and it will still look like a cube--this is where you need a subdivision surface, AKA “subD” modeling--not to be confused with regular subdividing. SubD modeling is kind of the polygon equivalent of modeling with NURBS since it’s good for smooth organic surfaces but doesn’t take as much computing power. Adding a subD modifier to an existing model will add subdivisions AND a smoothing algorithm, similar to the render smoothing, but in this case the actual geometry is smoothed.
Boolean Booling
Booling uses add, subtract and intersect functions to combine geometry in different ways. It’s an important feature for modeling and for 3D printing in particular. You can easily make a model by taking a bunch of primitives and sticking them together and it will look okay and render okay, but probably won’t print well--if at all.
The problem is mashed together parts have hidden geometry that is inside the other parts, but still gets sliced and printed. This can cause problems with the slicing software--making it take longer to finish, or even crash--and the resulting model will often be messed up, have holes or weird geometry problems.
Even if the model slices properly it will take longer to print and use more materials since all that hidden geometry is being printed as well. In the long run it’s far better to suck it up and combine all the parts into one shell using the boole function.
If boole operations are spitting out funky geometry, make sure neither object has typical geometry problems (see repair section at end of article). Good news for CAD users, you should get clean, problem-free booles almost every time.
Booling uses add, subtract and intersect functions to combine geometry in different ways. It’s an important feature for modeling and for 3D printing in particular. You can easily make a model by taking a bunch of primitives and sticking them together and it will look okay and render okay, but probably won’t print well--if at all.
The problem is mashed together parts have hidden geometry that is inside the other parts, but still gets sliced and printed. This can cause problems with the slicing software--making it take longer to finish, or even crash--and the resulting model will often be messed up, have holes or weird geometry problems.
Even if the model slices properly it will take longer to print and use more materials since all that hidden geometry is being printed as well. In the long run it’s far better to suck it up and combine all the parts into one shell using the boole function.
If boole operations are spitting out funky geometry, make sure neither object has typical geometry problems (see repair section at end of article). Good news for CAD users, you should get clean, problem-free booles almost every time.
Shelling
Shelling is mainly used to cut down on weight and save money when printing on machines that use a powder material such as 3DP, CJP or SLS. As you may recall, the powder printers use the loose powder in the print bed as support material. If the model is hollow, the support material from inside the model can be removed and you won’t be charged for it. In some programs you can simply use a hollow function, specify the wall thickness and blamo! it’s done. Unfortunately, most programs don’t have this function, so you have to do it manually by creating an inner shell.
First, you want to boole together most of the larger pieces which make up your model. This will create one continuous shell and if you’re lucky you can select the whole thing and extrude the surface inward, basically making a duplicate shell inside the original, like we did in the MintyBoost video. Unfortunately, this does not work for a lot of models, so keep a close eye on the inner shell when you try this. If geometry starts to intersect itself or freak out, it’s not going to work.
The more tedious option is to boole the entire model together to make one large shell and build the inner shell manually. This can be done by putting a simple primitive, such as a cube in the center of the model and carefully extruding and adjusting it outward to create a shell. It’s helpful to use the wireframe or x-ray mode in the side and top views so you can see inside. Remember that reference cube we made? Use it to eyeball the proper wall thickness for the shell, it doesn’t have to be pretty or even super precise since it’s on the inside and will never be seen. The last step is to combine that cube with the model and then cut escape holes for the support material.
INNER SHELL FOR JETCAR MADE FROM EXTRUDED CUBE.If you want to see what this gains you, upload the same model, one solid and one hollow, to Etba3ly3D and get a quote
This method won’t gain you anything on a home FFF printer since you can specify wall thickness (shells) and fill percentage in the slicing program. I will often have multiple versions of the same model tailored for specific printers.
Shelling is mainly used to cut down on weight and save money when printing on machines that use a powder material such as 3DP, CJP or SLS. As you may recall, the powder printers use the loose powder in the print bed as support material. If the model is hollow, the support material from inside the model can be removed and you won’t be charged for it. In some programs you can simply use a hollow function, specify the wall thickness and blamo! it’s done. Unfortunately, most programs don’t have this function, so you have to do it manually by creating an inner shell.
First, you want to boole together most of the larger pieces which make up your model. This will create one continuous shell and if you’re lucky you can select the whole thing and extrude the surface inward, basically making a duplicate shell inside the original, like we did in the MintyBoost video. Unfortunately, this does not work for a lot of models, so keep a close eye on the inner shell when you try this. If geometry starts to intersect itself or freak out, it’s not going to work.
The more tedious option is to boole the entire model together to make one large shell and build the inner shell manually. This can be done by putting a simple primitive, such as a cube in the center of the model and carefully extruding and adjusting it outward to create a shell. It’s helpful to use the wireframe or x-ray mode in the side and top views so you can see inside. Remember that reference cube we made? Use it to eyeball the proper wall thickness for the shell, it doesn’t have to be pretty or even super precise since it’s on the inside and will never be seen. The last step is to combine that cube with the model and then cut escape holes for the support material.
INNER SHELL FOR JETCAR MADE FROM EXTRUDED CUBE.If you want to see what this gains you, upload the same model, one solid and one hollow, to Etba3ly3D and get a quote
This method won’t gain you anything on a home FFF printer since you can specify wall thickness (shells) and fill percentage in the slicing program. I will often have multiple versions of the same model tailored for specific printers.
RSS Feed