3D Printed Drone Continued
Here it is, part 2 of my drone design adventure. If you haven’t had a chance, now is the time to check out part 1 here: Drone Racing Meets 3D Printing.
For a while now, I have been designing a drone frame that is to be 3D printed and that will be able to handle the abuse that comes with first-person view (FPV) drone racing. It needs to be tough enough not break into pieces if I come in contact with racetrack obstacles or “gates”. It needs to be strong enough to not break from high G-forces on high acceleration or sharp turns. It needs to be light enough to maximize agility and performance. Rigid enough to minimize mechanical noise, slim enough to minimize drag, simple enough to be 3D-printable by most, cool enough to interest people into the hobby, inexpensive enough, efficient enough, fast enough… (sigh)… You get the idea, we all want our own pet project to be the BEST!
Engineering and Design
I have been focusing on applying my engineering thought into my design of the drone frame. Consequently, I have gone through 5 iterations of the frame now. I think I found what I am looking for – or so I said the last 4 iterations. Using SOLIDWORKS Simulation I came to find an area that would be prone to fracture on small impact for each of the arms. I then decided to include a “raised perimeter” along the arms making the arms be more of a sideways I-Beam (see image below). Compressive stress concentrations are now focused along these raised perimeters which 3D printed Polycarbonate (Stratasys PC-10) can handle. I am glad that I have been changing my mind and kept correcting design issues while the drone was still in SOLIDWORKS. It is way better to find a design issue virtually than on the track.
Applying SOLIDWORKS Simulation
Motivated by the changes I was able to make after discovering them through SOLIDWORKS Simulation, I decided to continue my search for design issues. Specifically, I wanted to know more about the behavior of the drone frame under dynamic loading. Unnecessary vibrations, also called mechanical noise, are a really bad thing for a quadcopter that depends on sensors to function.
You may remember from childhood cartoons, a character singing really loud and hitting a specific note which shatters glass around the singer… Those notes are resonant frequencies of the glass. For my drone, the singer is the motors – they provide the vibrations. The frame is the glass from the cartoon example. Although the frame will not shatter with motor vibrations, it will vibrate more vigorously at those resonant frequencies. I need to know where they are in the range of the motor’s minimum and maximum RPM. Programmatically, I can then attenuate or “lower the volume” of the problematic resonant frequencies in the firmware of the flight controller board so that the sensors function well.
Running a Frequency Study
I used SOLIDWORKS Simulation again to run a frequency study. I found two resonant frequencies within the motor RPM range. One at 314.89 Hertz and another at 390.37 Hertz. Now some math (no worries its easy). I’ll spare you the easy math, but these are 18893.4 RPM and 23422.2 RPM respectively. These values are well within the max RPM of the motors I chose which can spin up to 30,000 RPM at full throttle. I will now update the firmware to attenuate these frequencies.
Next, I will focus on getting this drone 3D printed using our Stratasys 3D printers. I am sure I will learn a lot about the behavior of the frame at different printer settings. Stay tuned!