Throughout the course of this project, we have learnt many things about drone building and product prototyping. These are documented on this page.
DRONE BUILDING
Drone building is a delicate process requiring much precision and accuracy. It is not as simple as simply plugging a flight controller into a power board and connecting the motors – rather, much programming and calibration is necessary for the drone to be stable and able to fly.
KEY LEARNING POINTS
- Always check before soldering or gluing something
- Drone parts are fragile and delicate, so it is important to order extras in the event of mess-ups or crashes
- Calibration is not as easy as it seems
- Different parts from different brands may not be compatible with each other (one example: using the wrong length of screws which resulted in two smoking and burnt motors)
3D PRINTING
Working on this project allowed us to discover the endless possibilities of 3D printing. 3D printing can be used to custom print many objects that aren’t available for purchase.
For our payload’s container, we needed spacers attached to the container as well as a specific size that will fit the drone’s bottom plate. We also needed a lightweight container so as to not reduce the maximum load of the drone. Since it was nigh impossible to find a box that fit all these specifications, we designed a box in SolidWorks and had it printed.
KEY LEARNING POINTS
- How to use SolidWorks
- Designing a 3D object for additive manufacturing – the printer builds bottom-up, hence it is important to ensure that there is sufficient support for any protruding surfaces.
- Designing for tight fits – we needed a tight fit for the box and the lid, but the first box and lid we printed had too large a gap (0.5cm) between box and lid fitting. However, not leaving any leeway would have resulted in issues if the 3D printer was inaccurate. Thus, we designed a slanted box and lid fitting.
OTHERS
PLAN AND PREPARE ESPECIALLY FOR CONTINGENCY
It is important to plan in advance and order spare parts in case things don’t go our way. We had ordered parts too late and did not order extras, which led to an overly pressing schedule due to the lack of time to actually work with the parts once they had shipped.
This also led to issues later on during testing, when we did not have enough spare parts after a bad crash and had to immediately order extra motors, frames and propellers.
WORKS IN THEORY =/= ACTUALLY WORKS
Some planned parts looked good in our theoretical plan but did not actually work in practice.
For example, we had wanted to use a Pixhawk flight controller from the start due to Beverley’s one prior experience building a drone with that flight controller. Unfortunately, we had underestimated the difficulty of calibrating and programming Pixhawk flight controllers, and this proved infeasible in the end. The crash of the prototype 3 days before our final presentation further resulted in our need to use a manual F4 flight controller instead of the Pixhawk.
Another instance would be when we tried to use 6045 propellers (6 inches) on our mini 250 quadcopter. Though the 6045 propellers could be mounted without issues, when the propellers started rotating, they would be cutting all the wires in our drone. Thus, we had to change to 5045 propellers.
CHANGE AND ADAPT QUICKLY
Many things can and will go wrong during the course of a project. For example, shipping the parts took a long time. The drone kit that we ordered for our first prototype took more than 2 weeks to arrive, even after paying extra for express delivery, due to a public holiday in China. This slowed down our progress significantly.
Furthermore, usage of the drone kit proved infeasible for our project due to the weakness of the carbon fiber frame as well as the difficulty in programming the Radiolink Mini Pix hardware that came in the kit.
We had to change all these firmware quickly (from Radiolink Mini Pix to F4 flight controller and then to a HKPilot Pixhawk) to meet the deadlines for the project. In fact, just when we had managed to get the HKPilot Pixhawk working and flying, the prototype crashed, which resulted in the HKPilot Pixhawk being unusable. Thus we had to switch back to the F4.
KEEP AN OPEN MIND
A lot of the feedback we were given during the progress meetings as well as from our supervisor, Tony, was harsh, and it seemed like Dr Ho and friends were out to destroy our ideas and our project. However, ultimately this resulted in many improvements to our project as we had to think of different ways to address their issues with our product.