In collaboration with Urban Farmer Steakhouse in Portland, OR, our makerspace was given the opportunity to test out a new HD Glass filament that is advertised as transparent, "food and drink contact approved," and could print like PLA. Our first project was a Glencairn style whiskey glass to see how the filament performs, how well it can store liquid, and how transparent the material actually is. First, the 3D model was created in Autodesk Inventor (a Free student version can be downloaded here) and the right settings for the Makerbot Replicator 2 and wall thickness was found through trial and error. The .STL file can be downloaded below.
It turns out that there a balance between transparency and strength. If the wall thickness is very thin, it is very transparent, but weak. If the wall thickness is very thick, it is very strong, but not transparent. We found that the optimum balance is a wall thickness of 1.5mm, which is 3 shells thick, with the largest layer height possible of .4mm because our nozzle diameter is .4mm and the layer height cannot be greater than the nozzle diameter. We also found that the quality increases with a slower print speed and 35mm/s seemed to be a good print speed
Step 1: Results
First, I will admit that this is one of the best quality filament materials we have tested, but we did find that the claims of transparency may be a little over hyped. Our conclusion with the material is that it is possible to print very thin walls to get an almost glass-like appearance, but it was a very weak structure and couldn't hold liquid very well.
It seems to print best with the lowest quality settings, which does a thicker layer height of .4mm. This means that a keeping the material as thick and continuous as possible will result in better transparency. However, the Makerbot slicing program does a horrible job of creating the extrusion paths and if you look close, there is a huge seam down the side. Also, if the wall is too thin, it's not solid and there is an air gap between the outside and inside shell, which is not good for holding liquid as tiny holes can get filled.
In the bottom photos, the glass on the right is slightly more transparent because of a thicker layer height of .4mm. The layer height of the left glass is .3mm. We also noticed that with very thin walls, the 3 line segments that create the outside curve of the glass seem to be printing differently, which creates horizontal lines along the body. This effect could also be a result of the horrible Makerbot program. We could try some other slicing programs like Cura with the Replicator 2X and heated bed to see if we get better results, but the Makerbot has proprietary gcode that cannot accept sliced models from other programs.
As far as holding liquid, we came to the conclusion that this would be good for a one-time use as liquid can seep through micro holes in the material. In the last photo, we poured liquid in to the whiskey glass and it seems to hold for many hours without any issues as long as the optimal 1.5mm wall thickness was used with the largest layer height. However, after a few hours, the liquid did seem to find its way out eventually. We also tried using a heat gun for post processing of the glass wall, but it didn't seem to blend the layers of the glass wall, but it did seemed to strengthen the wall very slightly by making the wall shells stick a bit more.
Recommendations for improvements can be using a Stainless Steel nozzle, heated bed, and dedicated food only 3D printer for higher quality and food safe prints. This would be easier to clean and prevent contaminants from entering the walls of the glass.
Congratulations! You 3D printed a whiskey glass!