One, two, 3d print!
The final stretch
The previous steps on this journey can be found here and here!
As semester three dawned so did the endless worry of the 3D printing. I had now tried to print with DJCAD’s Digital Making and realised that the end result I was looking for was not going to be achieved with the current tools available. I wasn’t happy with the rough texture and visible printing lines that FDM printing provided, so I powered up Google and went to look for other solutions.
After some time it was clear that the needed detail I was looking for could perhaps only be achieved by using an ASL printer instead of FDM. SLA prints with resin, so it doesn’t leave behind such clear and visible layers as well as retaining so much more detail that it is mostly used for printing highly detailed miniatures. So pretty much what I need.
The fundamental problem with SLA printers is that DJCAD Digital Making does not have one. Well mostly it is a me problem, but a major problem nevertheless.
DJCAD Digital Making has several printers but they all are FDM printers, as seen below.
This required some correspondence, so I contacted the people from Digital Making and asked for their advice. As I had already seen from the website, they indeed didn’t have any SLA printers as they are quite messy and expensive to operate. The next best solution was either buy my own, use an outside firm or try to make the FDM work with different settings.
Since I’m a little hesitant to spend money for a machine of my own and since having an outside company print my stuff was also very expensive and time-consuming, we came to an agreement that we would try to alter the printing settings so that the nozzle used would be smallest they have and by lowering the layer height the obvious lines could possibly be less visible.
I also made sure to ask about the actual maximum limit for polygons per print and was given a number of 600k per print. This obviously is very little considering just the base model of my mermaid was more than 7 million. So the solution for this problem would be slicing the poor lady in smaller parts, so then the total for the entire model wouldn’t be 600k, but 600k per part then totalling multiple millions when glued together.
It means that at least the arms, torso, maybe even head and ears, tail and every other subtool would come in teeny bits and then I would need to figure out a way to connect them back together seamlessly. It would definitely be a lot of work afterwards, but it would give me more control over the quality of the parts and maybe even have less sanding in the end.
When I reached the point where I was somewhat done with the model I decided to do the first test print with these new settings we had discussed with the DJCAD staff. As I had planned, I axed my model into four parts; torso, tail (not in its entirety), spikes and hands. I decimated them all to be 600k each and exported them as STLs before sending the digital-making job request.
I asked for the torso to be 14cm tall, and then line the rest to that. Once printed and assembled the total height would be quite tall, so I’m excited to see how it turns out. The height in my mind would be pretty much the final, so after I finally have the model in front of me I can then decide if I need to go bigger or smaller. The sculpture needs to be big enough to be seen from a distance but not too big so that it’s too difficult to handle.
I’m also interested to see if I still need to slice the model even more, as currently her facial details are included with the torso’s polygon count. I might need to print the head by itself in the end if I need even more details.
3D printed dummy
The process of printing took a total of two weeks. That was way longer than I had anticipated, so it meant I’d need to start printing the final sculpture ASAP. I’d also need to ask the Digital Making if they could somehow prioritise my work and print faster. The time was seriously running out as I needed time for building, sanding, painting and whatnot afterwards too.
The dummy print was really good. And I was surprised. I had been so skeptical about the whole thing so I was pretty much gobsmacked the moment I collected the prints and how well they looked. There was some battle to take the supports out but other than that everything was pretty nice and smooth underneath. I’ll most likely print the head separately however, as the supports left a pretty nasty mark underneath her chin and was very difficult to clean up. Same story was seen underneath her breasts, but I was able to clean those pretty ok. The size of her was quite majestic as well, even though Sean told me to go even bigger for the final piece. Her scales printed very smoothly too, and only needed some light sanding underneath to clean them up. Overall, I was happy and would definitely print the final piece like this too.
As seen in the pictures the torso printed out pretty well, except the area underneath her chin. Some of the supports were super difficult to get off, so I had to trim the final bits with a dremel and then hope that I could be able to fix the damage by sanding. Same problem was seen on the back of her tail.
The biggest issue was the hands, however. The printing people had used support material that was supposed to melt away overnight if placed in water. I did so, but unfortunately the layers underneath of the supports were pretty horrific. There was a lot of deep damage to the hands so it took a lot of sanding and filling up with epoxy putty to level them out. I’m hoping one or two layers of plastic primer will even out the rest, but for sure I’ll be asking not to use this stuff for my final sculpt at all.
When it came the moment to attach all the pieces together I used two part superglue and epoxy putty (green stuff). This was something Duncan had suggested so I was happy to try. The process was pretty straightforward, just apply the activator on both pieces, wait for 60 seconds and then add bit of the superglue on one of them and hold for 30 seconds. Then I just had to leave the glued piece alone for 10 minutes and it was done. The bond was super strong (no pun intended) but since there was a visible seam I then added some green stuff on top and let it dry for 24 hours. I’ll then sand the pieces well and prime the whole piece with rustolenum plastic primer before painting.
The spikes printed out very well too, they were only wonky and bent since I didn’t spent too much time modelling them for testing purposes. The big question was what I could use in between them to mimic the fin flesh of a fish. I wanted the material to be translucent if possible, so I tried some normal tissues and cut out of a freezer bag. The tissue was extremely awkward to work with and left a nasty looking texture on top, so I might go with the plastic in the end. This might of course change as I paint the piece and it’s highly possible that the texture might end up being more interesting than the other.
I attached both side fin tests on the body with the same technique as with the other parts and waited to see the results after priming.
One thing I’ll switch for final print, however, is that I’ll most likely slice the spikes in half so that the fin process would be easier. There’s also the possibility that I could just sculpt the spikes out of epoxy putty straight away and skip the printing part.
And then there was the hair. For 3D I’ll of course just use hair geometry nodes, but for printing it was a whole another story. I had thought about just printing it as a huge mass and then covering that mass or diying the hair like in doll repaint videos. I wanted to try the yarn technique first, so I took some yarn and tried to brush it out with my metal brush like in those doll videos. Soon after starting the process I realised I had a serious problem I should’ve seen coming. I would not be able to brush out a long piece of hair like I wanted, as how yarn is made is by brushing smaller pieces together, before spinning it to a longer snare. So of course, when I brushed it out I would be left with a massively shorter version than what I started with.
There is in fact an answer to this problem from another doll maker, and I was pretty happy to try this until came the next step: straighten the hair with your straightener. I don’t own a straightener. So the question was, should I get one, or use the same money and buy an actual wig and use the fake hair from that to make my own mini wig? It would also mean that I don’t have to prep the hair at all but just simply cut it, attach it and style it. After discussion with the staff (and voicing some concerns of using ready-made materials) I came to the conclusion that it would be way easier and less time-consuming to just buy hair extensions/doll hair/wig hair and use that. It would take away the prep time of yarn hair and I can glue it on right away.
After the epoxy putty hardened I took the sculpture outside for sanding. At the package it said it wouldn’t be good to inhale the dried epoxy, so this was something I had to do. Unfortunately I managed to pick a day with worst possible weather, so as I was sanding away it actually started to rain. I knew I really had to get the tests done asap so I ended up not sanding the sculpture completely, just enough to see how easily the putty was to sand and then quickly put together the ugliest looking paper tent for spray painting the primer.
Little did I know that my idea of protecting the print from rain was actually pretty counter productive, as thanks to the wind the tent was ready for take off right away. This obviously resulted some pretty bad spraying conditions and the primer landed on the print VERY badly. So badly in fact, that I even tried to wipe the first layer off and start again which wasn’t really happening.
The end result was pretty sad, but at least I learned the hard way that do not use any kind of paper tents or other light obstructive elements in the Scottish weather. I even ended up painting my balcony railing and the ground white. Next time I’ll just lay down a sheet of fabric and make sure to spray the primer from a good distance away, just like the can actually instructed me to do.
Ignoring the obvious surface imperfections of the finished product, I am still confident with my decision with using the primer. The white colour did indeed cover the green of the putty as intended, and after making sure that the final version will be sanded 100% well the seams will be practically invisible.
I also discovered how much I hated the fins, so I decided to give up on the idea of transparency in the end. It would be so much more effort to either tape all the transparent fins before priming and painting, or attaching them in the end. Considering they won’t be that visible underneath all the resin anyways I figured that the battle would be wasted time and there’s a lot of other more pressing matters to attend to. I changed the plan and ended up actually modelling the fins too to be printed.
Final version of the sculpt goes to print
At this point, I was finally ready to retopo and rig the model, so I exported the sculpture from ZBrush to Blender. The process of that can be found here.
Before taking the model back to ZBrush I took a look at the images again I took of the life model. I ended up going for the “shy maiden” feel we explored and from that I made sketches for the layout. I looked some reference for the tail from snake coils to drawings of mermaids and then combined those all to one drawing.
After the whole painful process of making sure that the digital side of the character will work fine I exported the posed rig back into Zbrush as an OBJ. Since the rig wasn’t exactly the most delicate and perfectly working I had to do some sculpting to smooth out wrinkles and odd angles. It was pretty easy and only took me one day to do this. I added the scales back onto the tail (small individual ones nearly taking my sanity with them), carved indents for the eyes and added a sneaky detail behind her ears. As said before, I decided to use the printed fins after all, as the posing of them would’ve been an absolute nightmare without any guide. Instead of printing the spikes like before I’ll just build those with the putty. Easier that way.
I sliced the model into several subtools and exported them all as STLs like before. I have to admit, I did make a small mistake with the hands as I feel like a better choice would’ve been to slice off the fingers too, especially on the left hand. This was unfortunately realised after I sent the files for printing, so we’ll just have to see what happens and work with what we get!
The printing will take again two weeks, and then we’ll see what the end result is.
Prepping the model
It indeed took a good while, but I was finally able to collect my prints. This happened very last minute, as there were some personal problems back home that affected my working schedule tremendously. Only a week before the deadline I travelled to uni to collect my prints and was in awe. The sheer size of the prints was majestic! I opted for three centimeters more in torso height, compared to the 14cm I had in my test print. This definitely made a huge difference, as the totalling heigh towered over 30cm.
I took the prints home and started the literally painstaking process of cleaning up the supports. This alone took me five full days. I had no idea why the supports on her tail were so stuck on, with only very small pieces coming off at the time. My hands were swollen and in pain after spending hours after hours picking the pieces with my pliers. After the fifth day came to an end I decided that there just simply wasn’t a way to take off all of them, no matter how much time I spent on the process. Thankfully the extra material was indeed on her tail, and since it will be submerged underneath the resin I could try to sand it to my best ability and then hope that the tinted resin covers the rest. I could also try to play into the texture and pretend that it’s moss gathered on top of her tail after years and years being underwater.
On top of this texture issue, I also noticed back at home that one of the tail fins was missing. I tried frantically emailing and calling the Digital Making, to only finally get an reply on Tuesday (deadline being on Friday) that their logs indicate that the fin was printed but nowhere to be found. There is a possibility that it was accidentally discarded as trash. The guys at the lab started to reprint the fin right away though, but there is no possibility that the fin will be ready by assessment.
I was already extremely sad with myself that I wasn’t able to finish the sculpture by the dealine, but now not even have it glued in one piece was even more frustrating. There is time until the Masters Show, so I hope that the fin finishes quickly so I can start preparing for that.
For the assessment I decided to build the model to best of my ability and then with 2D magic adding the plan for the future. Below there’s a slideshow to show the planned progress of the build:
