EOS 3D
EOS 3D
This document tracks progress for the YSWS portion of this project. It also serves as a journal documenting the development process.
Preface
Before starting this project, I must outline the key requirements to ensure clear project direction and successful execution. I initially came up with the following requirements:
- $300 (USD) budget
- Klipper compatibility
- No calibration printing is required
- 180x180mm build space
- Cantilever structure
- CFD-optimized part cooling
- Safety considerations:
- Must not pose a fire hazard
- Must not cause harm to the user
- PicoMMU compatibility
Although I have extensive experience working with 3D printers, this is my first time designing one. Through this process, I aim to deepen my understanding of 3D printer mechanics, refine my CAD skills, and improve my approach to mechanical design. Additionally, thanks to Hack Club, this project offers the added benefit of a free printer.
The name of this project, EOS 3D, is inspired by the Greek goddess of dawn, Eos. Dawn symbolizes new beginnings, reflecting a personal milestone in my projects and an entry point for tinkerers.
Day 1 (February 1st, 2025)
Hours worked: 8
Total hours on project: 8
BOM Selection
Creating the Bill of Materials (BOM) required balancing cost-effectiveness with performance and safety. While selecting parts, I took into account the following:
- A budget hotend from a reputable brand to ensure safety and good flow.
- A reliable direct-drive extruder.
- The Cartographer probe for its high-resolution bed meshing capabilities and automatic Z-offset calibration via Carto Survey.
- BTT SKR Mini and a Raspberry PI for its cost-effective control.
- A cheap heated bed with a PEI spring steel sheet.
- A PSU with a large overhead from a reputable brand.
Compiling and refining the BOM took approximately four hours.
Toolhead Development
With the BOM completed, I decided to tackle the printer’s tool head first, as it serves as an “anchor point” for the build.
I began by modelling the X-axis rods and bearings to establish clearance constraints between the toolhead and the rods.
Next, I imported the CAD models for the Sherpa Mini extruder and the E3D V6 hotend, then started designing a mounting solution. I opted for a clamping mechanism for the hotend, as shown in the image below:
By the end of the day, I refined the mounting system and securely integrated the Sherpa Mini with the mount. Below is the final design on day 1 (ignore the mesh issues on the Sherpa; those were present in the CAD model):
Day 2 (February 2nd, 2025)
Hours worked: 2
Total hours on project: 10
Further Toolhead Development
Today, I completed the connection of the toolhead to the linear bearings with inspiration from the Prusa Mini. I began modeling the cooling fan duct but encountered challenges in designing one that is both functional and aesthetically pleasing.
Here is the final model of the day:
Day 3 (February 3rd, 2025)
Hours worked: 1
Total hours on project: 11
Fan Development
Today, I did minimal work; however, I finished and placed the part cooling fan. Tomorrow, I will begin designing it and running CFD tests.
Day 4 & 5 (February 5th-6th, 2025)
Hours worked: 4
Total hours on project: 15
First Fan Iteration Completed
After extensive trial and error, I have an initial design for the fan duct. The goal was to implement a dual exhaust setup, which presents more challenges but should enable more even cooling across the print.
Next Steps Before CFD Analysis:
- Raise the outlets so they are positioned above the nozzle.
- Adjust the outlet angles to prevent airflow from hitting the nozzle.
- Increase clearance to the hotend to prevent outlet deformation
With these changes, I plan to start optimizing the design tomorrow before running CFD simulations. Below are images of the fan duct (rear view) and the complete model in its current form.
Day 6 (February 7th, 2025)
Hours worked: 5
Total hours on project: 20
Computational Fluid Dynamics
With the latest changes in place, it was finally time to put the design through CFD analysis. Having never used CFD software before, I knew this would be a challenging process.
I initially tried Autodesk CFD and ANSYS Workbench, but getting them to work felt overwhelming. Eventually, I discovered SimScale, which turned out to super user-friendly. I followed this YouTube tutorial to set everything up.
After a long computation wait, I finally got to see the airflow results—and they were really bad. Not entirely surprising for an initial design, but definitely something to improve. The results are shown below:
Tomorrow, I’ll begin optimizing the design and rerun the CFD analysis to see how much improvement can be made.
Day 7 (February 8th, 2025)
Hours worked: 6
Total hours on project: 26
Finishing the Part Cooling Duct
After reflecting on yesterday’s analysis, I decided to try a different ducting style to improve overall airflow. This new design features a 180° duct, which I hoped would distribute air more evenly. I started by sketching the outlet, connecting it to the rest of the duct, and running it through CFD. After setting everything up, I obtained the following result:
CFD Run 1
As shown in the results, the airflow was heavily biased to the left, causing a lack of airflow through the right slit and the exit airflow being angled.
For the next iteration, I attempted to make the turn more gradual to prevent the airflow from favoring the left.
CFD Run 2
The results looked good! However, there were still three issues:
- The airflow was still left-biased.
- There was a lot of extra space (look at the top left and bottom right of the ducts).
- There was no airflow through the rightmost duct when there was no object “printing,” as shown below. This is the big one.
For the next run, I made some adjustments to the ducting to make it more smooth and prevent the sharp corner. I also made the throat leading up to the outlet smaller and added a fillet to the outermost outlets.
CFD Run 3
With the throat being tighter, I was hoping for the air to be forced into the right side; however, this was not the case.
I then made the interior of the outlet slightly smaller horizontally so there was less open space and made some adjustments to the ducting leading up to the outlet.
CFD Run 4
These changes were a big step forward in resolving the overall bias issue. There was still some airflow lacking on the right side; however, the airflow was less skewed.
For the next run, I added a barrier down the middle to help balance the airflow between the two halves.
CFD Run 5
Adding the barrier finally solved the issue, and there is really nice flow. It pretty much goes straight down the middle, and each slit is getting sufficient airflow. I decided this was good enough and moved on.
The only thing I would really change is the angle of the slits, but this design is sufficient and looks nicer.
Post CFD
With the part cooling fan optimized, I quickly designed a mount for everything and called it a night.
Note
Before I end this journal entry, I want to mention that I did use AI (specifically ChatGPT) to assist me in optimizing the design. I have no formal education in this space and no mentors to help me out, so using AI was my best bet.
Tomorrow, I will be finishing off the toolhead with the mount for the Cartographer and PicoMMU “Hub.”
Day 8 (February 10th, 2025)
Hours worked: 2
Total hours on project: 28
Finishing the toolhead
Today, I modified the part cooling fan mount to improve its aesthetics and structural integrity. I also installed the Cartographer probe using a simple mount. Tomorrow, I will begin working on the X-axis!
Note: The PicoMMU infrastructure will be added after the initial prototype is completed.
