The Biomass-Fungi project aims to eventually utilize waste materials to cultivate algae, which will then used to propagate a particular fungus strain through some chosen powder. As the fungi grows, it binds to the particles, and if we can control the shape in which the fungi can grow, we can create 3D objects, i.e. 3D printing. As such, the part that concerns me, the mechanical engineering major, is the part where we need to automate the process of laying down anti-fungal binder in a very particular way to create a 3D shell for the fungus to grow into.

The project is still very much in the “proof-of-concept” stage, and as such, the PhD student(s) working on this project for the last 3 years have developed a sort-of 3D printed build plate mechanism, while relying on having a person manually lay down each layer of powder, and manually dispensing “binder” via pipette.

This manual method, however, has resulted in wildly inconsistent results, so my work involves developing a machine that will produce consistent parts with adjustable parameters (such as layer height, compaction, binder density, etc.). As such, I am essentially developing a custom binder-jet 3D printer that can be sterilized.

As a sort of introduction to the project, the main Ph.D student presented me with a small mock-up made from some steel brackets that he epoxied together, and asked me to devise a way to make a plate go up & down.

So I sketched out a plan on my tablet, 3D modeled the components in Solidworks, then used the Rapid Prototyping Studio to manufacture my parts all within 1 day. I 3D printed most of the components, and plasma cut the build plate itself.

The final assembly for V1 was very rough, and only had about 1 inch of total Z travel, and the plate had a lot of wiggle room to both rotate and translate around. But with this aside, I still gained insight and used it to talk about my design and learn more about the needs of the project. By this point, all I knew was that I was told to make the steel plate go up and down, without much more context.

By now I was learning more and more about the project, and now I knew that it needed to be steralized, preferably in an autoclave (blasted with hot, pressurized steam), to prevent contamination.

My biggest goal with version 2 would be to reduce the amount of space between the bed and the surrounding box. As such, I decided to manufacture the box via sheet metal bending, with the idea that the corners could be sealed, and the walls will turn out fairly perpendicular.

I spent about 6 hours sketching out an initial design, originally starting with a 2D cross section view of the design, thinking about where I would place certain components. I included some rough dimensions to sort of plan things out, and make sure things would fit reasonably. After this, it only took me about 4 hours to then translate that into a full 3D assembly in solidworks, using McMaster Carr to get the models of the hardware I wanted to use, and 3D modeling my own custom components for everything else.

I presented my findings to Dr. Pei as well as the other PhD students on the team, and everyone approved and appreciated the idea, so I was given permission to order the parts I needed in order to create a prototype. As of 8/31/25, I am still waiting on the most important part, the stainless steel sheet metal, which is largely blocking me from making progress, because I want to finalize the job box before making other parts (like the bed, or the rubber gasket seal).

From the highest level, BJT printing works by gluing layers of powder together to form a solid 3D object. Typically the binder is dispensed by a printhead utilizing the piezoelectric effect (whereby certain crystals will flex under electrical impulses) onto whatever powder is being used. With modern BJT printers from companies like StrataSYS or ExOne, these powders can range from Titanum, Tungsten, Ceramic, Aluminum, or any combination of those to achieve unique material properties.

The general process is as follows

• Lower print bed by 1 layer height (∂z, typically 0.1 mm)
• “Refresh” a new powder layer at a thickness ∂z
• Lay down binder in specified 2D pattern for that layer
• Repeat

Providing powder for each new layer is typically done in one of two ways:

• Using a sifter / roller mechanism to drop powder onto the print bed from a hopper, then rolling it smooth
• Having two job boxes with a roller (spreader) mechanism, so that when the print bed moves down by ∂z, the other plate moves up by ∂z, so that the volume of powder provided is enough to fill the volume of the new layer.

For the fungi printer, the powder we are using will either be Hemp-hurd, or coarse sawdust. In either case, we found sifting to be very inconsistent, jamming the sieve very frequently. As such I chose to implement the 2nd option of having dual boxes, where one plate goes up, and the other goes down. This way, the powder simply just needs to be pushed across the plate.

The other components will be explained in better detail in the next few paragraphs.

After I presented my V2 design to Dr. Pei, I was informed that the scope of the project would eventually include automating the other parts of the part creation process (i.e. laying down a layer of new powder, and dropping binder in a certain pattern).

I knew that a complete overhaul of V2 would be needed, since I never intended the sheet metal job box to be incorporated into a larger machine. The thin sheet metal meant that I couldn’t easily mount it into a larger mechamism.

So I started from scratch. I found a few open source Binder Jet 3D printers that I could reference from. (INSERT LINKS) Project Oasis, Plan B, and (THE SLS ONE) were all used for inspiration as far as the design and form of the printer. I utilized the gantry from the Voron 0.2 project as a relatively simple implementation of the CoreXY movement system, and expanded the frame to allow for the spreader mechanism as well as mounting space for the job box.

reword that to better explain the parts of a binderjet 3d printer holy shit

im goin to bed dawg…

This part was a bit of a struggle for me. I got information about the project in bits & pieces, being guided through the process. Looking back, I could have extrapolated some things and probably figured it out sooner, but I eventually got enough information and context to start on the V3 design.

• At the beginning for V1, the problem was pretty vague and solution-oriented.
• For V2, I took the same prompt, and expanded on it, asking questions on the way
• For V3, I finally had enough understanding of the project and what they needed to be able to start developing the full printer

At first, the more I thought about the problem, the more daunting the task seemed. After collecting myself I identified 4 main sub-problems and used that as a starting place:

• Z axis
• X/Y axis (Gantry)
• Roller
• Binder deposition

During the last step, I started by remodeling the assembly from the Voron 0.2 project in Solidworks.

Then I drafted a roller system that would fit within the frame. I ran into a lot of issues with finding a place for the roller motor to go where it wouldn’t hit the frame. As of 9/5 I haven’t had the time to fix this quite yet, but I may need to change the width of the frame, or simply reduce the size of the job box (which I would rather not do)

The third part I tackled was designing the job box system, which especially gave me a lot of trouble at the start, so I had to break it down by drafting a pros/cons list for different methods of manufacturing the box itself. My first thought was to make one monolothic structure using casting techniques, since I could 3D print a sacrificial job box, then use sand casting to get what I want. This idea ended up being scrapped due to the poor surface quality and the difficulty in machining the bottom corner surfaces. What I opted for instead

Lastly, the final part was mostly independent of the other systems, and I researched a peristaltic pump that would fit on the frame of the printer, with a hose that went all the way to the printhead on the gantry with a nozzle to control droplet size.