Open comments for this post
CERES MK2: Devlog 10
What’s New: I made a couple of changes to the CAD design (which I will showcase in the next devlog), and I also began soldering the motors to the wires in preparation of setting them up with the motor drivers. As the 3d printed parts start getting ready, I’ve been attaching them to the aluminum frame, as seen below
Currently, the biggest issue with the assembly of the robot is the difficulty that comes with ensuring that the T-slot nuts are inserted and held in the exact position necessary to mount the body panels. I attempted to compensate for this earlier in the design process by designing v-slot liners, but they did not work properly as they were not able to hold the T-slot nuts effectively.
What’s Next: The next step in the project is to work on setting up the Raspberry Pi, which I have not yet done. I also want to test out the ESP32 with the motors and Joy-Con soon.
Open comments for this post
CERES MK2: Devlog 9
What’s New: I designed a Raspberry Pi camera module mount and housing, in order to accommodate the image recognition software on the raspberry Pi. I’ve begun to print out all the CAD components, so I’ll be able to add onto the (already assembled) aluminum frame soon. I’ve also begun to work on a rudimentary system to control the weed robot via a Nintendo Switch Joy-con in order to test its capabilities before paring it with the Raspberry pi.
To use the Joy-Con with the esp32, I used a library called BluePads32, and it allows for a connection between an esp32 and any controller. It was relatively easy to set up due to there being in-depth documentation on the library. You can visit the documentation here if you’d like to try it yourself!
What’s Next: The next step in the project, while the 3d printed components are printing, is to start soldering and wiring up the motors to the motor drivers, and start experimenting with serial communication between the ESP and Raspberry Pi
Open comments for this post
CERES MK2: Devlog 8
What’s New: Today, I designed multiple new body panels on the robot in order to properly house the electronics from the elements. In addition, I’ve begun the physical construction of the robot’s frame, which I’ve attached below
The biggest hurdle today was assembling the frame for sure. Because I was using sliding v-rail T-nuts instead of the insertable type, I would have to partially dismantle the frame every time I have to add any panels or mounts.
What’s Next: I WILL start working on the code for the ESP32 next time, as I’ve been putting it off for a while now. I’ll also start physically wiring up the ESP32 with the placeholder L298n drivers so that I can test the robot.
Open comments for this post
CERES MK2: Devlog 7
What’s New: Because I completely overhauled the design of the bot (Mk2), I started re-designing the lower casing that protects the electronics and motors from the elements! I also used this opportunity to redesign the way the two separate plates (of the casing) fit together in order to allow for both easier assembly and better overall durability.
Some Other Things: I had to redesign the wheels as well to accomodate the new 5mm threaded rods that I plan to use for the axles. I also changed up the color scheme plan - I’m going for a more modern, high tech aesthetic.
What’s Next: I have yet to implement the electronics into the Mk2, and the design is purely mechanical right now. I have to design the mounts for the RPI, esp32, motor drivers, and etc. I also want to start designing the PCB for the bot so that the motor drivers are integrated directly with the ESP32.
Open comments for this post
CERES MK2: Devlog 6
What’s New: I decided to completely overhaul the design of the robot and essentially start from scratch. In the initial iteration, the entire weight of the robot was resting on the motor’s shaft/axle. In this new design, I’ve pivoted towards a gearbox design that will allow for the weight of the robot to rest on 4 axles instead of the motor itself. So far, I have only designed the gearbox itself as well as the mounts for the axles to sit on. This ensures the longevity of the robot as the wheels are no longer directly attached to the motors.
The gearbox itself uses two gears, with a 1:1 gear ratio as I did not want to change the RPM output of the wheel, as I felt that the 100rpm provided by the motors was sufficient in terms of both torque and speed. The gearbox mounts to the frame using 4 sliding T-nuts and M5 screws. The axle itself is an M5 threaded rod, and the rod sits on 3 ball bearings with an inner diameter of 5mm to rotate freely.
What’s Next: I’ll have to redesign a majority of the body panels, as well as the wheels themselves. I’ll likely do that before I move onto designing new mounts for the electronics.
Open comments for this post
Whats new: Designed and implemented V-slot rail sliders for the robot frame. This is done for easier assembly, as it no longer requires the precise positioning the the T-nuts when assembling and mounting body panels for the frame. I also made some small aesthetic design adjustments
I faced some issues with the alignment of the lower body panel and the frame. For some reason, they’re not aligning even though I believe all my measurements are correct. I have yet to resolve the issue, so I’ll have to save that for next time.
Whats Next: The CAD/frame/body panels are done for the most part. There are some issues (like mentioned above) that still need to be fixed, and there are a few adjustments I still want to make. The next step will be to move on to the code for the ESP32, so that the robot can be controlled manually. After that, I’ll start designing a PCB to integrate the components better.
Open comments for this post
I designed custom v-slot liners for the aluminum extrusions, allowing for easier assembly of the body panels.
Open comments for this post
I altered the locations where the motor drivers and raspberry pi are mounted so that they have more protection from the weather. I also fixed the issues with the tire mirroring (which can be seen in the previous iterations) and I designed the top cover to allow for the electronics to be more accessible.
Open comments for this post
I’ve been printing out and assembling the wheels and tires of the Weedkiller bot. The tires are printed in 95A TPU. I’ve also been re-doing the layout of the electronics on the robot, which I’ll post later.
Open comments for this post
I’ve updated the color scheme of the Weed-killing robot to look more agricultural. I also finished desigining the mounting holes for the lower housing plate and designed mounts for the raspberry pi and motor drivers as well. I’ll start off by using L298n drivers and I hope to eventually start using my own PCB that incorporates an esp32 module and a H-bridge motor driver.
Open comments for this post
Say hello to the Weedkiller bot! I’ve yet to name it, but I’m currently working on its outer shell/casing so that it can be weatherproof.