3D Printing in Practice

3D Printing

For the full research into 3D printing history and its uses see my previous post over here!

To continue the story of my 3D printing journey, it was time to start thinking of how I could use it to my advance best. I would be printing a lot of small items and trinkets on a relatively small scale so I needed first to understand the limits that I was facing to understand further what things are possible and what isn’t. I had already successfully 3D printed the Moomin character I modelled last semester, and was quite happy to move onwards. I faced a few problems with that model, mainly the quality of the model wasn’t high enough and also since the supports were automatically added, removing them resulted in breakage and unevenness (or was it the impatient artist who couldn’t wait to see the end result, who knows).

This meant that I had to be a little more careful when modelling and really think about the end result and how printable it could be. I wasn’t too afraid of the character model, which should be pretty straightforward, but I was more cautious about the vegetation and small accessories. How easy would it be to print and clean those? I also had to think about the hair of the character, as I had designed it to be wet and curly. How annoying would it be to sand down afterwards? This path of discovery would be a very expensive one-way road to trial and error indeed!

Small details

I had to start somewhere, so I ended up modelling some examples of 3D prints I wanted to send to DJCAD for printing. I decided to create a good variety of options so that I would have enough material to study and alternate the designs if needed. The main question marks to me were the vegetation, hair, tail texture and small accessories. I created the first batch of these elements in ZBrush and then sent them over to printing, hoping that they would all turn out okay. The models weren’t by any means finished or anatomically correct, I was mainly interested in the texture and the quality of the print. Miss Groke had some serious problems with the surface and the weird lines that appeared thanks to the model being too low quality. I checked the polygons later on and it looks like she was around 1 million polygons, which was clearly not enough. This time I opted for over 5 million, which definitely started to slow my software a little.

I started with the most obvious issue which was the tail scales and spikes. I mostly used alphas to speed up the process, with varying levels of height. I also left two options for the tail fins, one with the “fin flesh” in between and one with just the structural spikes. Since I would like the flesh between the spikes to be transparent, I’d need to use fabric or something similar afterwards, but the problem is that I don’t know how the spikes would print alone. After thinking of this issue for a while I had a look through DJCAD digital printing page and saw that there were in fact several different printing materials to choose from, as well as colours.

I chose to print five prints in total, four in black PLA (which was the same cheapest material I used for the Groke in white) and one in transparent PET-G. I sent the form over and waited a few days for a confirmation. When it came the person on the other end told me that the files were so big that they were unable to even open them. This was definitely not what I had planned as my idea was to up the polygons to avoid the uneven surface of my first attempt last semester. I opened the files again and lowered the polygons to around 1 million each and uploaded them again for printing. It took two days and the prints were ready for collection. I went to pick them up a little scared as I wasn’t sure if the end result would be what I had hoped for.

The pair of spiky friends were the first ones I turned my attention to. For sure, again, the polycount was way too low as the result was pretty lined and rough. Interestingly the PET-G was even worse to wear than the PLA one. At collection, it was full of wisps of thin plastic and the thinner spikes were super rough. Ignoring the bad sides I was positively surprised about the thin flesh parts in between. The effect was very nice, not too transparent but not too opaque. It would not be wise to print the entire thing in PET-G, but I could print only the fleshy parts and glue the on afterwards. I could even engrave shallow grooves on the base of the fin to make the bond between the fin and tail stronger. I would need to try printing with more polygons again, hoping they’d pass the first inspection.

The second pair was the planes with alpha scales. I tried out multiple alphas for different results to try to decide what kind of scales I would like my print to have. Again, too few polygons but the result wasn’t too bad. I could still make out the details and the height differences well. I was drawn towards the taller scales as they would be more visible from afar and on top of it they actually felt a little sharp too which would go with the idea of spiky fish.

For the last print, I made a doughnut with the same scale alphas (my favourites). I wanted to try the scales on a curved surface and to my delight this print was the best of the bunch. Too bad I made it too small, it would’ve been the best self-defence bracelet. Again, I liked the tallest scales the most, especially the really spiky ones were actually very sharp so they really seemed a little dangerous. For further tests, I’m going to try a mixture of scales in a random pattern.

I was pretty happy about the results, they definitely gave me insight into how I can model and print the tail in the most believable way. I had hoped I would have managed to print more during this semester, but I ended up being way too tangled with the going live module. From day one of semester three I will start a serious factory of print tests to make sure that the final print will be the best possible.