Dimension FAIL

The first lab in ME 498N (Advanced Topics in Rapid Prototyping) had my team trying to identify limitations or failure modes for the Stratasys Dimension 3D Printer. This system is a 3-axis gantry fused deposition modeler; in other words, a small bead of plastic is extruded to "print" a part layer by layer. An image from SolidWorks and an image of the actual part from our first attempt, which was designed by my team member, are shown below.

This part was designed to test several specifications of the Dimension. The columns on either end were made to find the minimum column thickness for square and circular columns; we were also attempting to determine if the height of the column changed the results. The curved portions in the center of the part were to test how the machine handled vertical columns that transitioned to horizontal columns. From this model we were able to determine basic information about minimum column thickness. Most of the features did not provide us with data because they were covered in support material (support material is created by the machine's software prior to printing; if the plastic is extruded over air with no support material it will sag). The support material is designed to be removed after the part is built, but because our features were so small the part would break if we were to remove the support material. From our first attempt we learned that the Dimension is not able to build very small vertical columns and will not even attempt to do so. Using this information, I designed the second model which is shown below.

This model was built to trick the Dimension software into not building support material where it should. The base has beams coming off at 15, 30, 45, 60, and 75 degrees above horizontal. Each beam has thin columns extruded down to the bottom of the model. We assumed the Dimension software would use the columns in the design as support material, and since we knew from the first design that it would not attempt to build columns this small, we would be able to test the ability of the machine to build overhangs, instead of relying on the specification that is given in the manual. Unfortunately this model did not work because the software could not handle the file size (12 MB). The reason the file was so large is because the file type that is used to print (.STL) uses triangles to approximate the entire geometry; circular columns require a lot of triangles for a close approximation. The image below shows the second version which replaced all of the circular columns with square versions. The spacing between columns was also increased so fewer would be needed.

The reason the beam with an angle of 75 degrees above horizontal does not have columns below it is because we learned that the machine is able to build features at an angle without support material if they have an angle greater than around 45 degrees. The beams have different heights because they were designed to have the same horizontal overhang distance. After importing this model into the Dimension software we were able to determine that it would not attempt to build support material below any of the beams. The final printed part is shown below.

As shown in the image above, we were successful in tricking the machine into printing without support material. The beams built at 15 and 30 degrees show that material started to sag as the part was built. From what we had learned about the machine's ability to build at an angle, we were not surprised that the beams at 30 degrees and below had problems while the beams built at 45 degrees and above did not. We were not able to explain why the Dimension built base support material below the columns since it did not attempt to build our columns or support material on top of the base support material. The professor for this class, Mark Ganter, submitted our part to Stratasys; we were successful in finding and exploiting a software problem that tricked one of their high end machines.

Team member: Jacob Hiester