Have you ever wondered what the world looks like from the perspective of a frisbee? What we initially thought would be a straightforward project became several months of experimentation and prototyping leading up to Maker Faire 2013. If you haven’t seen FrisbeeCam before, check out the video from our booth last year.
We had a blast sharing FrisbeeCam with fellow makers last year and have been hard at work improving it ever since. As with the initial project, we’ve had to continually experiment and prototype to improve all aspects of the device — overall usability, durability, and the quality of the captured footage. In addition, we gained valuable experience in switching our prototyping almost entirely to 3d printing — as we’ve learned, it’s not as simple as print and forget.
While the prototypes that we showed at Maker Faire 2013 utilized several 3D-printed parts, at the time we had only recently started to experiment with the technology. In the intervening year we’ve built a lot more experience with additive prototyping methods, and we can say in retrospect that the investment in learning the quirks of a single printer has proven crucial in shortening our prototyping cycles. Take a look at the photos below for more details.
With regards to the usability of the device, we knew that the large vertical fin we had in 2013 both impaired the throwing motions with the disc and added significantly to the overall weight. As a first step we 3D-printed a base chassis that could hold foam fins of widely varying sizes. After several iterations of testing and experimentation, we’ve converged on the current design, which represents a significant size and weight reduction from last year’s device. We’ve also experimented with a variety of mounting strategies, bearing designs, and adhesives to significantly improve its durability in everyday use.
While the video footage from last year’s prototype was sufficient to convey the potential of this concept, it lacked the necessary stability and raw quality to be used in the HD videos that serve as the baseline format for socially shared content today. Drawing from the many resources available from the aerial RC community (links below) we tested new camera modules that have dramatically higher quality from before.
We hope that you enjoyed seeing FrisbeeCam at Maker Faire and look forward to presenting all of these changes at the upcoming Bay Area Faire! Take a look at the photos and captions below to check out many of the things that we’ve learned as makers this past year.
As last year’s May show date approached, we were scrambling (like all good makers) to build our cleanest and most polished prototype yet. Up until that point, our best working prototype was milled out of HDPE plastic and quite heavy to throw. By the time we debuted at Maker Faire 2013, we had managed to create two prototypes that were 3D printed, but they were still very rough and not sturdy enough for regular use.
Over the last year we’ve improved our ability to model using free 3D design tools like SketchUp and Tinkercad, and we have been able to significantly improve the sturdiness of the design as well as make more repeatable prints. We’ve also managed to find a sweet spot between what desktop 3D printing can do and what we can do with off the shelf parts. Desktop 3D printing, for instance, does not have the resolution to support printing high quality threaded bolts. Instead, we make holes for which off-the-shelf nylon bolts can fit into without having to additionally drill holes, eliminating yet another step in assembly.
We experimented with having the fin manufactured together with the chassis since, in theory, having a single body and fewer parts to put together should result in more a stable flight and consistency across devices. Instead we learned that the solid plastic fin added several significant disadvantages to the design. First, solid plastic fins added more weight to the overall apparatus compared to the foam fins we had before. Second, because the fin was now a part of the chassis, if the fin got caught on anything (like, say, the sidewalk…) the whole chassis would be compromised. Third, we found that the stiffness of the fin actually detracted from the stability of the camera footage. The softer foam fins appeared to provide a bit of dampening to any sudden changes in cross winds possibly because the fin itself could bend.
After breaking a plastic chassis and fin that took weeks to get back from Shapeways, we decided to go back to foam fins, that were cheap and easy to replace. One big change to the way the foam fins are attached from last year’s design is that no glue or adhesive is required. The new chassis has a split in the tail that allows for a fin to friction-fit into. An additional hole was added to the design to provide alignment along the center of the chassis. We nicknamed this new design prototype the Sparrow.
Our first attempt of mounting the camera to the frisbee used an under-frisbee bearing mount. The idea was that the style of mount would reduce obstructions to the airflow going across the surface of the frisbee. The downside to this approach was that it required more complex assembly because you had to cut a centered hole in the frisbee to accommodate the rotating shaft, and it required a heavier mounting scheme, using a screw that would hold the bearing mount to the underside of the frisbee.
This year, as we focused more on 3D printing all the parts, we evaluated simplifying the assembly process and made everything top-side mounted. The results proved that the reduced weight of the newer style mount had a more effective impact than the airflow obstruction. With that said, we made efforts to raise the bulk of the bearing mount as high as we could off the top surface of the frisbee.
This new style mount would make a rigid mount between the inner race of the bearing and the bottom side of the mount. Prototypes of the mount were adhered to the top surface of the frisbee using double-sided sticky tape. The top chassis was a friction mount to the outer race of the bearing. We used blue painters’ tape to provide the friction, which worked well.
Lessons learned in this exercise were: The area of the bottom mount where the double-sided sticky tape was used needed to be quite large. When the frisbee comes in contact with a hard surface upon landing, that would cause a lot of torque and it would be pulled apart. The base of the mount has an area that is used to hold/embed the hex head of the bolt-axle. Initially we had a 90-degree edge to this, as if the bolt head was stretched through the base. During higher torque landings, the base would tend to break right at that corner. Later revisions used a slanted edge, making the mass at the corner greater, giving it more strength.
Last year we started with a bare PCB 720p 30fps #18. It had the advantage of being very lightweight, 12g including battery. This year we experimented with several other cameras that had higher resolution. The results were dramatic, but at the expense of being much heavier, ~30g. The camera we are working with now is the 1080p, 30fps Mobius action cam.
Footage from our new, improved FrisbeeCam: