3D Modeling & 3D Printing

We were asked to design and 3D print a joint that fits 3 to 4 wooden rods/profile. Thus, I got the idea to design this vase-liked joint structure.

The first and the most important step for me was to gain some inspiration and try to figure out what most of these joints look like. Those images above reflect how these joints mainly looks like and how they functions to connect sticks or holders.

Finally, I determined to model the shape like the branches, with one main branch and four minor branches towards four directions.

The modeling process were mainly composed of three functions in Rhino, which are Pipe, Fillet Edge and Trim.

However, as we were informed of the size of the round wooden rod, which has a diameter of 20 mm, we need to make the model size a bit larger. As a result, the inner diameter is about 20.04mm, while the outer diameter is about 24 mm, since we cannot make the wall of the model too thin.

Also, at first I tried to select all parts of the model and trim them simultaneously. However, since the inner construction of the model will become really complicated if doing the trim together, I later subtract them step by step.

Though I intended to use Fillet Surface to smoothly connect the pipes, there turned out to make tiny gaps between the joints of those pipes. And the final printed version was not so satisfying, since there hardly exists any strong connections between those pipes.

Finally, I had to give up my initial trail on building the model and accepted what Professor Andy suggested, namely first edit connecting parts based on columns and later use subtract to cope with the inner part. This turns out to be a better way as I don't need to struggle with the overlapping part of the connections, which caused my previous model to be open.

However, there also existed some other problems for this time. The Fillet Function caused some unexpected results, namely some twisted surfaces inside the pipes. The problems could even not be fixed with Meshmixer. As a result, this trail ended up in failure.

This time, Professor Andy suggested another way to build the model, namely using curves, then surfaces, finally constructing solids to approach what we intend to achieve.

Here, there's a quite convenient tool called ArrayPolar, which easily helps create symmetric shapes around a pointed center. Then we can build surfaces based on these closed curves, and finally join those surfaces together to construct a solid.

Finally, we use the Subtract tool to make the inside space.

The model has to be a close shape to make it printable. Also, since my model is a little bit not so stable, it will need extra support to keep its balance. What's worth mentioning is that we could flip the whole model to make it face down to limit the support printing material.

Then we transmit the model to the

Firstly, we can adjust the Infill of the model to adjust the inside support structure, either to make the grids smaller or larger. How intense these grids range depend on what model we print.

We are also able to preview the slice of the model to check the whole printing process.

Since most 3D printing process takes relatively a long period, it's not possible for users to stand by to watch the whole process. As a result, it's worth mentioning that we need to pay much attention to check the printing of the first layer, considering it's the foundation of later printing.

Finally, it's time to insert the model file into the SD card of the printer. Actually, some tinny mistakes do not cause big problems, however, if there appear to exist some mess on the bottom, that will cause big problems.

Since the first several layers looks quite similar to the preview version in the application, I could leave and pick the printed model the next day.

The next day, I went back to check the model. The 3D printer has already completed its work. Since the melting material would become a little sticky when printing, we need to use the shovel to shovel the model off the plane.

Also, since the supporting part has thinner structure than the main part, we could easily take that part down to get the final joint model.

The most important point must be, never pay much hope to the Fillet tool to deal with surfaces, as most times it will built overlapping surfaces with unsolvable problems, though the model itself might seem to be quite good.

Also, the tool called Arraypolar is really impressive as it helps to save much time to create symmetric structure.

In addition, we need to be aware of the tiny differences between the printed model and our digital model. As a result, during the constructing process, we need to slightly enlarge the inner diameter of the model than the diameter of the wooden rod.