Skip to content

Latest commit

 

History

History
71 lines (42 loc) · 3.24 KB

logbook.md

File metadata and controls

71 lines (42 loc) · 3.24 KB
Raw

Logbook (inverse chronological order)

Development Checklist for single interphalangeal joint

I'm currently focused on getting one joint completely built, controllable, and characterized.

1. Mechanical Design
  • Implement basic gearbox.
  • Test out basic gearbox.
  • Implement custom rotated transmission for compactness.
  • Test out custom rotated transmission.
  • Implement clean design for finger housing (and easy assembly).
  • Compose the interphalangeal joints to make 2 joints in series.
2. Control System Design
  • Derive MPC controller for impedance control via current sensing.
  • Formalize MPC controller design in a small whitepaper.
  • Calibrate MPC controller gains.
  • Test out MPC controller on simple servo.
  • Test out MPC controller on custom interphalangeal joint.
3. Hardware Design
  • Implement basic end-to-end circuit (shunt resistor for current, encoder, microcontroller).
  • Decide if I want to centralize all of the electronics and control or if they should be partially distributed via CAN bus for easier wiring.
  • Design PCBs.
  • Get PCBs made.

May 18th, 2024

I designed the MPC controller for impedance control (stiffness/damping) using current sensing! I even wrote up a whitepaper that walks through the derivation and has a psuedocode section for algorithm implementation.

I annotated some of the flaws of the MK6 design below, and have already addressed a few in the new gearbox:

May 5, 2024

Updated the repo readme, added milestones, assets, logbook, and motivations. Also, drafted a schematic for the transmission that will allow me to rotate the motor 90 degrees to be inline with the rest of the joint without jutting out:

Older than May 5, 2024

I am just summarizing these, since I didn't write logs for some of the sparse earlier work I did here.

First, I created a bare-bones test to recreate the servo gearbox from scratch. The goal here was to prove that I can replicate the gearbox in a parameterized manner, such that I can integrate it into the body of the finger for compactness.

Then, I built a simple enclosure around it, since the gearbox was snagging due to plate flexture.

Then, I experimented a bit with magnetic encoder placement.

Then, I added an end-effector to get a better feel for the motor torque.

Then, I did some more work around the encoder to better conform to the narrow specifications for the distance between the magnet and the encoder itself. I also connected the encoder to the end-effector to be able to couple their movements for easier testing.