Design Process

The best choice based off of our drone’s requirements would be an octocopter. The octocopter drones are able to lift the most weight and can sustain flight even after losing a motor. This provides what amounts to built-in redundancy, which is good for safety. That being said, battery life will be constantly strained due to the increased power requirements of eight large motors. We came to this final design from the decisions made first semester with the Morph Chart, which can be found here.

We also need to redesign the octocopter’s undercarriage so that it can properly hold the payload. The space between the two legs is currently not wide enough to support a large pizza and the drop mechanism. For the construction of the drop mechanism, we plan to use aluminum as the main material. This material is both cheap and durable. However, it does weigh significantly more than plastic or carbon fiber, both materials that are slightly harder to manufacture. We chose an octocopter in order to support the weight of our payload.

tarot-x8-big006     simple-octocopter-outline

The size of the drone needed to account for the size of a pizza box as well; the props needed to have as little (preferably zero) overlap with the box to avoid a large downdraft. Shown in the left figure above, the drone is designed to carry professional camera and gimbal systems with a payload capability exceeding the weight of over two pizzas. By not using up the full payload capacity, we can expect flight times to be close to the manufacturer estimate. The top right figure shows the footprint of each of the propellers as well as the body of the drone. The red dotted square represents a 14″ large pizza. By not having the pizza sit directly underneath the draft of the propellers, we can avoid turbulent air underneath the propellers influence flight dynamics and efficiency.

Assuming the total weight of the drone with payload to be around 8 kg, or 17.64 lbs, each of the 8 motors must provide a thrust capable of lifting 1 kg. At this level, each motor will run at about 5.2 Amps.

The radio telemetry requires 0.525 Amps while the flight controller takes roughly 0.36 Amps. The GPS draws around 0.022 Amps. In total, the electronics draw roughly 42.5 Amps.

Our current battery is rated at 16,000 mA h. So 16,000 mA h / 42,500 mA = 0.376 hours. We multiply this value by 0.7 to make allowances for external factors and inefficiency. So we approximate an expected battery life of 15.8 minutes at maximum output.

With this knowledge we performed a series of flight tests to confirm these data points, but did not get very far.

It is not if you break something on a drone, but rather WHEN something gets broken. For testing purposes, the “Broken Arrow” quadcopter was born. “Broken Arrow” was designed to be a rapid-fix chassis so we wouldn’t spend as much down time waiting for parts to come in. Instead we chose to produce it out of things that were cheap and easy to replace due to access of in-house machining tools. “Broken Arrow” was designed so that we could test the Ardupilot flight controller’s replacement, the TI MSP432 LaunchPad, without worrying about inevitable crashes.

We 3D-printed parts designed in CAD to be used for the motor mounts and arm braces. We then utilized the laser cutter to cut out the main body of the copter. In the center of the drone we have two layers of 3mm plywood with electronics attached to each. At each corner of the centerpieces we have a 3D printed piece attaching the main body to the carbon fiber arrow arms with two rods per arm along each of the top and bottom plates. At the end of the arm is a 3D printed piece that attaches the motor to the arm. This drone is significantly less power hungry, so it only requires a 2-3S LiPo battery. The battery life ranges from 9 to 13 minutes, allowing ample time for testing. For an exact list of either drones’ parts and their specs, please follow this link.

The primary coding development process will take place in Code Composer Studio using the language, C. For guidance, we consulted TI’s E2E forums. The Github for this project can be found here.