Marble Machine Prototype

 

This post will describe the printer settings, the CAD models printed, and the results for our initial prototypes. We initially printed all of the parts that comprised The 3D Printed Marble Machine #2 which was a total of 4 parts as seen in the CAD blog post. The prototype for the marble machine consisted of the track and cylinder (attached), the screw, and the two screw caps.

 

The largest print was the marble track. As mentioned in the previous post, an Ultimaker 3 printer was used and the build material was PLA and no support material was used. The parameters used were a layer height of 0.25 mm, an infill density of 20%, a grid infill pattern, and a print speed of 100 mm/s. The track was oriented in Cura with the length of the base along the x-axis, the width of the base along the y-axis, and the build direction along the z-axis. The track took 15 hours and 58 minutes to print but had good quality on the first print attempt. The only surface defects present occurred at sharper edges of the marble track which do not impact the functionality of the machine. More specifically, there were some loose strands of material that could be removed by simply pulling off. Therefore there was very little post processing required. Given the good quality  and relatively short print time (See Print Parameters blog post), these are the settings that will be used in the future. The track prototype is seen in Figure 1.

 

Figure 1. The initial track prototype (note: it is a completed assembly in this image, the track is only the turquoise part).

 

img_20170417_154927575_hdr

 

The next parts that were printed were the two screw cap pieces. Their CAD files are seen in Figures 2 and 3. These parts were quite simple compared to the rest of the parts that comprise the marble machine and not nearly as important. Both parts were printed at the same time using an Ultimaker 3 printer and gray PLA build material and no support structure. The parameters used to print the parts were a layer height of 0.25 mm, an infill density of 20%, a grid infill pattern, and a print speed of 100 mm/s. The print time was 1 hour and 2 minutes in order to print both parts. When examining the prototypes, they had great quality and no noticeable defects. For these reasons, we decided to use these as part of the final Marble Machine.

Figures 2 and 3. The CAD files for the two screw cap components.

Figure 4. The printed screw cap components. Not all of the brim was removed at the time this pictures was taken and the gap is meant to show there are two parts present.

screw-caps

 

One of the most important prototypes that was printed was the screw. This part needed to have high quality as it is used to transport the marbles from the base of the track to the top. It was also one of the most challenging parts to print given its unique geometry. One of the most important factors we discussed was what orientation we should print it in. If we printed it on its side as seen in Figure 5, the screw would most likely get deformed along the edge in contact with the plate and not be able to transport the ball properly. Therefore, we decided to print the screw in the vertical direction as seen in Figure 6. Since it is a tall part, we were worried that the printer head could knock it over as the height increases. In order to mitigate this risk, we increased the brim width from 7 mm to 14 mm. Since the quality of the screw was very important, we made sure to use small layer heights and slowed the print speed. The screw was printed on an Ultimaker 3 printer and used gray PLA build material and no support. We did not want to include support between each screw thread as it could incorporate deformities when the support is removed. The parameters used to print the screw were a layer height of 0.1 mm, an infill density of 20%, a grid infill pattern, and a print speed of 70 mm/s. Using these parameters, the extra brim, and the vertical orientation resulted in a print time of 2 hours and 46 minutes. When examining the prototype see in Figure 7, it seemed to print the screw very well. The small layer height allowed for a smooth surface. When we tried fitting the screw on the peg that attaches it to the track, it was too tight of a fit. With some post processing (sanding) we were able to enlarge the hole on the bottom of the screw enough to allow for smooth motion. Another problem was one small section of the screw (back right when looking at the Ultimaker) seemed to have a deformity along the length of it. We noticed during printing that a fan or some sort of blowing air was pushing the edge of the screw up in this section. It is barely noticeable in the final part and when we tested it, the screw was capable of transporting the marble from the bottom of the track to the top of the track (although it was not as easy as expected). Since the screw was functional despite its small deformation, we decided to use this for the final Marble Machine.

Figures 5 and 6. The horizontal screw orientation (left) and vertical screw orientation (right).

screw-orientation vertical-screw

 

Figure 7. The screw prototype (assembled with cap pieces).

img_20170417_155436250_hdr
As mentioned previously, it is a little bit difficult to turn the screw when the marble is traveling up it. While we initially believed it was due to the screw, we now think it is due to the size of the marble. After problem solving the issue for some time it was determined that there was a lack of clearance between the screw and cylinder geometry at certain portions of the lift. The original print on Thingiverse specified using a 9.5-10 mm steel ball bearing with the track. After taking measurements with a digital caliper, it was determined that the current steel ball bearing we were using was 10 mm in diameter. In an attempt to resolve this issue, we printed a 9.5 mm PLA ball for the marble track. In spite of post-processing efforts, the ball was rough around the edges and was not a perfect circle. Additionally, it did not weight enough to make it all the way down the track, despite printing it with an infill density of 100%. However, the cylinder lift binding issue was no longer an issue suggesting the marble size was indeed the problem. Due to the poor quality and weight issues, the printed marble will no longer be continued to use.
Finally, the last prototype we printed was a Polycarbonate (PC) cylinder.  After discussions with Professor Rudolph and Nikhil, the material with the highest potential for printing transparent was PC. We thought it would be interesting to create a transparent cylinder so that you could see the marble rise up the lift. We decided to try and print just the cylinder (see Figure 8 for the CAD model) with PC for transparency. We used an Ultimaker 3 printer with PC as the build material. The parameters used were a layer height of 0.25 mm, an infill density of 20%, a grid infill pattern, and a print speed of 100 mm/s. Another parameter that was changed was the nozzle temp. Normally, PLA is printed with a nozzle temp of 190 degrees celsius; however, the PC was printed at a temperature of 250 degrees celsius. Additionally, the bed temp was turned up as high as it would go, but this was far short of the expected temp that should be used for printing PC. Unfortunately, the print did not turn out correctly towards the top portion of the cylinder as seen in Figure 9 and the promised transparency was not present as seen in Figure 10. Instead, the material refracted light in such a way after re-solidifying that made it difficult to see anything in or outside of the lift area. While a single layer of PC may be transparent, the cylinder we printed had a couple of layers of thickness. This attempt for improvement gave us insight that transparency can not be achieved for our cylinder printed with PC. Given these issues, a PC cylinder will not be used for the final design.
Figure 8. The Cylinder CAD file.
Figures 9 and 10. The PC cylinder prototype with the deformities at the top and lack of transparency evident.
unnamed unnamed-1