factorio.com
Friday Facts #447 - Factorio that you can touch
Hello,
we have very unusual Friday facts for you.
* * *
* * *
### Journey to the physical world Pard
It all started in the hot summer of 2024 when we worked really hard to finish the Space Age DLC. The game was almost ready, but we wanted some feedback from our players. A decision was made to organise a playtesting event (FFF-427), where teams of players could play the game from start to finish and provide us with in-person feedback and criticism to deliver the best product you all deserved.
Since we are all geeks and love to tinker, we asked our friends at Prusa Research to join us and create some accompanying programmes and presentations for all attendees.
Being fans of Factorio themselves, they agreed to join, and the plan was put into motion. When the event started, we realised that it would be cool to print some Factorio models right at the event, so Fearghall quickly created a simple model of the Wriggler from everyone's favourite Gleba.
With the conclusion of our event, everyone returned to work on Factorio, but we all wanted to return to the idea of Factorio models in the real world. In the months following the release and subsequent bug-fixing and polishing period, we started working on some prototypes in our spare time. Several models were made, printed, and painted, but they lacked a concept. We let it sit for a while and brainstormed what would be the best direction. The answer was actually very simple: transport belts.
Belts are the core of every factory from the early hours to the very late game. Jarosław started designing the concept of a grid system where you could place every entity closely following the game. The first version of inserters soon followed. During this phase, we were printing countless prototypes. It soon became clear that we needed two versions of almost every model. One set of models is created with high precision, allowing push-fit assembly, and the other is made with high clearance in mind for people who want to glue everything permanently or have clearance issues with their printers.
While Jarosław was busy creating one player-built entity after the other, Fearghall started working on enemies. The first iteration of the biter spawner soon followed. Then came chests, followed by the player character, then a stone furnace… Our model set grew nicely, but we knew that we had to draw a line somewhere, because trying to release every Factorio entity would take too much time. With that in mind, we focused on the early game. We ended up with 15 model sets containing 65 individual models in 247 STL files.
* * *
1/17
Testing how far we can push the limits of model size.
2/17
Painting of the first enemy prototype to ensure that models will look good both in plain plastic and fully painted...
3/17
...Finished demolisher. The model was later scrapped because it wasn't really reflecting in-game visuals.
4/17
Several models in early stages of development. Belts were tested and fine-tuned for clearance issues. Enemies needed major updates.
5/17
Painting the biter spawner to determine whether the fully finished model adequately represents the original vision for the game...
6/17
...The biter spawner is now fully finished. Unfortunately, we realized that substantial changes are still needed.
7/17
Most entities went through a series of iterations. Countless print runs were associated with this process.
8/16
First test print of the biter. It was later resized, and it received many improvements for a smooth printing experience.
9/17
Fine-tuning of printer limits regarding potential model details on the spitter.
10/17
The final version of the spitter, finding enemies outside the world of Factorio.
11/17
Test print of the late version of the electric mining drill with a coat of silver primer.
12/17
Our catalogue of models was starting to gain momentum. At this point, we were confident that a public release was possible.
13/17
The biter spawner finally received a model update. It still wasn't finished, but we were getting much closer to the finish line.
14/17
The belt lane balancer is painted with model grass added. It was clear that complete dioramas are possible.
15/17
One of the earlier tests of walls and their integration into the grid system.
16/17
Testing how far we can push the limits of model sizes in the opposite direction. Joint modification was needed.
17/17
Finally, all the models are finished. Time to take some glamour shots.
* * *
### Why 3D Printing? Pard
In the modern digital world, there is a decrease in physical memorabilia. Physical releases of games are now almost a relic of the past. We always thought about having something for fans and ourselves as a reminder of times that we spent with the game. We at Wube are a software company, not a toy manufacturer, and we have the best community in the world. Creating some expensive (limited) collector's items didn't sit right with us. We have decided to go with 3D models as a little "thank you for your support" gift so anyone can download them and print them for themselves.
A second equally important reason is that we love your creativity, and we want to give you a chance to iterate on things we created and give them your own spin. We can't wait to see it.
* * *
### Reimagining of Factorio Jarosław
I will not lie, I slept on 3D printing for a while. It's probably for the best - I do have a passion for miniature-based board games and background as a 3D artist. So, when I finally took the risk and decided to look into it - well, let's say neither of my 3D printers has seen much downtime since then. Whether it's miniatures, card organizers, inserts or wall-high paint-stands, there is always something new to print. And of course, in case you are wondering, that does not make my "pile of shame" any smaller.
With this premise, I know you can imagine why my ears perked up when the first hints of 3D printed models for Factorio started circulating in the office. Little did I know I was soon to partake in one of the most fascinating endeavours I ever… well, partook in.
If you are a regular in these here parts, you will likely know all our entities start their life as 3D models. That might make it seem like translating them into prints would be an easy job - in all truth, I thought so, and I of all people should know better! The journey turned out to be a bit more complex than I originally imagined!
The most obvious problem concerning FDM printing is the issue of supports. If you are not well versed in how these printers work, you can imagine building a roof with LEGO bricks. An angled roof is easy - you just offset each next brick by a single row, and there is your roof. If you wanted to make a perfectly flat one, though? The bricks do not connect on the sides, so each next one would need to float in the air… unless there was something to support them from underneath. For our models, we wanted to avoid extra supports wherever possible - they can be visually distracting when made to stay, otherwise they can be a bother to remove. As such, part of my work was to find ways to separate our models into pieces that can be printed without supports and then connected into full models. Considering I am a fan of scale models and gunplas, it was an incredible experience to create my own!
Now, the other problem is a bit different in nature. How would I say it…? Factorio models tend to be really, really crazy (in the best possible way). Do you know the adage "what you see is what you get"? Well, yeah, no. Not at all applicable in this case!
Factorio's isometric view gives a lot of space to experiment and play, and my fellow artists used that opportunity mercilessly. Sometimes what you see is, in fact, merely a collection of free-floating objects that just happen to look like a machine when looked at from the exact correct perspective. It rarely gets quite that bad, but the perspective is often a problem. There is a huge discrepancy between what the machines are and how they feel - all a function of the isometric view mixed with some creative light-and-shadow tricks. If you were to take a measuring tape to the mining drill for example, you might discover the North-South and the West-East variants are not quite identical - they do look like it, though! That problem goes double for height of the entities - in game, it is often restricted to fit in the grid-space of one tile, yet made to look much taller than that suggests. That created a unique and somewhat contradictory problem: designing 3D models that looked like Factorio models look, and not how they are.
There is one more issue recreating models made for an isometric game - and that is the dark spaces. Just like the dark side of the Moon, the back-sides of our machines are filled with horrors and sleeping eldritch-gods. Do not worry, though - I did replace them with freshly imagined, safe machinery :)
Partaking in this journey was surely one of the most interesting challenges in my career. I would love to thank Wube and Prusa for making it possible - as well as for the enthusiasm they shared! I honestly do hope more studios decide to share their work in such a way (looking at you, certain team of astronauts!)
Special thanks go to my colleague, Pard, for his insatiable hunger to test my models. May thy filament never runeth dry, good sir!
* * *
### Designing enemies Fearghall
While Jarosław toiled away hand crafting belts and assemblers and failing his "lazy bastard" run in the process, I set about the task of bringing the "organic" entities to life. Thankfully for me, I was able to avoid the pain of the isometric perspective because most of the organic entities have fully rotating spritesheets, and so the models are complete from most angles without too many tricks of the camera.
However, the models were still designed for pre-rendering only, and as such had some unique challenges. For example, the biter spawner was covered by an invisible skin, and almost all of the detail in the in-game images was confined to the texture data - something that cannot be transferred to a 3D printable file. However, after a week or so of cleaning up and baking out the model, it now prints very well with no support or issues.
It also became immediately obvious that most of the organic prints could not be made support-free like Jarosław has done so well with all the machines. The geometry is simply too complex, and many of the parts would be far too small to print and attach together. I still endeavoured to make them easily printed, however. The idle engineer pose, for example, has printed without supports many times, and the worm poses (aside from the 90-degree attacking pose) print entirely without support. But alas, there are some things which are simply impossible to print without support, such as the running pose.
One of the main issues with the character and biter family was the damage left by the support interface. The surface where the print and support meet can be quite ugly, so I was keen that none of the models should require so much support as to impact the visual quality.
With the engineer it was quite simple, the backpack needed to be split off, and there needed to be a slot and pin added, but otherwise a slight angle was all that was needed for decent printing.
With the biter family however, there was a little more work. There was the big issue of the legs: they would absolutely not have printed in the "correct" orientation, and splitting in the centre of the biter and leaving a large seam was an absolute non-option. So the only sensible solution was to print them upside down! Simply printing them like this as a single part would likely have worked, but would have left horrible scarring on the most visible surface of the model, so I decided the shell needed to become its own part so it could be correctly oriented. This also had the benefit of meaning the shells are easily printed in the correct colour for the biter size!
With that in mind, I could arrange all the biter legs to require as little support as possible in the new orientation, and merged a few of the legs for added strength - bearing in mind that they would need to stand on their own legs in reality. In order to make the shells attach, I sculpted in some extra reliefs and ensured both surfaces had just the right level of clearance. The final result has almost no visible support interfacing, and a very satisfying "snap!"
* * *
### Where to download? Pard
All models are available to download right now from our Prusa Brands profile on Printables. Special thanks go to Štěpán Feik, who helped us make all of this happen. From joining us at our playtesting event to helping us cross the finish line this morning, he has been part of the journey throughout. You can read his take on our collaboration here.
* * *
As always, let us know what you print at the usual places.
