I am building a 3 Degree of Freedom arm. This is in an attempt to better understand engineering mechanics and advance my knowledge of robotics and hopefully machine learning. I am to be able to learn inverse kinematics and basics of design through this.
Hello again guys and today we have another devlog. Sorry it’s been a while since my last devlog but I have a lot of exciting stuff to show.
Law of cosines was a bit annoying to learn and to apply considering my method is finnicky but I managed to get all my servos to move into position based on real time, and my turntable also works!
I severely underanticipated the amount of power this would draw, and my psu can provide 9v1a stepped down to 5v wheras my arm needs over 5v6.5a. In order to solve this, I will get a PD trigger board which I will hook up to a PCA9865 module in order to power this. I also learned a ton about SDA and SCL signalling, which I thought was like CAN but I was completely wrong
Although we still do not have a webend for me to run machine learning on this, I am able to manipulate all my servos.
One of the problem with the power draw is that my servos will jitter and release certain items. In order to circumvent this I will design an object and claw similar to Samples in the FTC game Into The Deep for 2024-25. This way I have something to dig into
Next up I have to work on a webend that way I do not have to constantly be plugged into my ESP-32
I have to wire up my PCA9865 (the SCL SDA will save me so so many wires)
And lastly implement machine learning! Adios my friends!
Yesterday my P2s was delivered! I’m excited to use it and I have begun printing some parts.
One thing I did not think about when designeing my parts is tolerences, especially for 3d printing. Today I spent some time on OnShape redesigning 2 of my parts for printing. More parts to follow, but this is is finally being built!
Unfortunately I have not ordered my servos yet (although I do have my stepper motor) so we will have to wait on Amazon for that.
Tomorrows devlog should have some parts assembled.
Thanks for reading!
Today I begun the wiring for it. Since I am still awaiting my amazon order for parts and my 3d printer, I found a tool called Cirkit Designer where I could wire my parts up, and program a few of them.
Unfortunately, my ULN2003 driver does not have simulation, but atleast I can wire it so that I have a reference for my program. This way, I know what parts to order
Alongside this, I began programming this. I am coding this in VSCode with Pymakr, that way I can code this in a language I am already familiar with instead of having to switch to C++, which I have never used.
I will continue to work on first controlling the servos and motors raw using an arduino I have lying around, but I will try to get my true code done soon.
I finished building my first prototype of the arm. This is going to likely be my base for the whole project until I am able to buy parts and a 3d printer for this project.
Currently here is the configuration
Base is driven by a 28BYJ-48 Stepper motor
Shoulder is driven by a MG996R
Elbow Wrist and Claw are driven by MG90S
This should hopefully lead to a maximum range of 10-11 inches with a max load of around 200-250 grams.
Hopefully, I will be able to power this with an ESP32-S3 in order to run machine learning algorithms and inverse kinematics that way I can identify and pick up objects autonomously.
