Today, we got the temperature harnesses made and hooked up to the battery management system. The BMS uses the temperature and voltage information about each cell to ensure safety of the battery pack during both charging and discharging. The next step will be to cycle the battery pack a few times from fully charged to fully discharged to balance all the cells.
We integrated the vision classifier algorithms with the controller for the arm. The video shows the stereo cameras finding and locating the pre-cached sample and moving the arm to point at the sample. There is still work to do to improve the stereo cameras and depth perception, but at least this is a start!
We received our GoPro a few days ago! Hopefully we will be able to make more cool time-lapses as the project goes on.
Our batteries are finally in. The fiberglass angle extrusion holds the batteries snugly against the I-beam. A little Hysol 120HP holds the fiberglass extrusion to the chassis members. A piece of dense polyurethane foam keeps the batteries from sliding around horizontally. We carefully connected each series pack of 8 cells to the main contactor and ground blocks.
We set up shop in the hallway so that we can start collecting vision and lidar data to test our algorithms. Our test rig “Polly” is far from our final design, but gets the job done.
Our Tesla came back from EKWB with a waterblock mounted! This is a custom full size waterblock specifically for the K20. It is milled to ensure that it makes proper contact with every chip that needs cooling. One huge advantage over the air cooler is that it makes the card only 1 slot wide. This frees up a PCI-E slot on our motherboard for the parallel port GPIO card.
Yesterday, we used Hysol 9430 two part epoxy and glass microspheres to structurally bond the fiberglass I-beam in the robot. This provides lateral stiffness for the robot. We used the glass microspheres to provide structure to the epoxy. By adding a small fillet between the fiberglass and the side panels of the robot, we significantly increase the strength of the joint.
We assembled our low voltage cutoff circuit. This is a small little circuit that monitors the battery pack voltage. If the voltage falls below 22.5V (or ~2.8V per cell), the cutoff circuit will open the contactor and prevent damage to the batteries. We have cutoff set a little bit higher than recommended by the data sheet, but we decided to have a little more safety margin because the battery cells are expensive.
Our Harmonic Drive actuator for the camera mast came in today! This is a pretty cool actuator with 10,000 count absolute encoder, 100:1 Harmonic gearbox, and AC motor all in one nice little package.
We wanted to use a 750W DC-input ATX power supply from a previous robot built in the lab. Considering the abuse it had taken over the years and the general lack of quality in the PCB, we decided to Hi-Pot test it to 1kV. Unfortunately, the insulation failed somewhere and the PSU did not pass the Hi-Pot test. Time to find a new power supply!