The original arliss design began as a senior design project. After I enrolled in graduate school my advisor asked if I wanted to participate in arliss for real and go to Nevada. I gladly accepted under his stipulation that my team changed the design.
The original senior design used a Raspberry Pie A+, an Xbee with SparkFun XBee Explorer USB, an Adafruit GPS, a Sparkfun breadboard power supply, and a usb wifi dongle.
Part of our design requirements, besides all the difficulties that come with arliss like driving and navigating, was to incorporate some other function. We decided on creating a jumping mechanism that would provide an escape from ruts or other obstacles.
This proved to be incredibly challenging, but it was done and worked well enough. The jumping would have never escaped any type of hindrance, however it lifted off the ground which looked impressive—and looks are all that matter. We all got A’s for our efforts.We were very concerned with clearing obstacles, but as we later found out this was not a very significant problem
That summer I met with my new team and we discussed a new design. In the meeting we all agreed the jumping leg had to go, however it needed to be replaced with something. After a few hours of research we found a very interesting passive wheel transformation paper1 from the Seoul National University and decided to repurpose the idea.
The designs were drawn up and cut out of aluminum with a waterjet. These initial wheels were very overweight but they allowed us to refine the drawings.
We decided to cut the final wheels out of a sheet of carbon fiber. This proved to be much lighter, however, the material delaminated in a few places. The carbon fiber wheels were also not very smooth to open and close. They moved nothing like what was shown in the video. Regardless, we continued with the design of the chassis. Initially, we 3D printed a simple flat chassis to test the code on; pictured is the assembled rover. Because the rover wheels were larger than the largest width of the chassis we needed stabilizer legs to prevent rollovers. Numerous arliss designs utilize metal tape measure as a strong but flexible stabilizer, so we incorporated the idea.
The next iteration of the chassis was again 3D printed and split in half; the idea was to glue foam on either half and sandwich the electronics inside. This was an interesting design choice that made the electronics extremely secure and helped with wire management. But, the design also made accessing the electronics a nightmare, as we discovered.
In order to release the parachute from the rover, a parachute release mechanism was designed. A lot of previous rovers used nichrome wire to burn the parachute attachment wires so we incorporated the idea. Once the rover realizes it has stopped falling a relay will trip connecting the nichrome to the battery and burn the connecting string. The parachute release was the most consistently working addition to the rover. If all else failed, we knew the parachute was going to get released.
After finalizing the design of the rover, there was one more test we begrudgingly had to perform—the drop test. So, one day before we left the rover was wrapped up, parachute attached, and thrown off the parking tower. We had all the replacement parts, but nothing prepares you for the harsh reality of seeing all your work broken.
The damage was not too bad. The flexible wheel couplers bent and the motor mount snapped off. So, we machined new couplers and bolted instead of glued the motor mount. We also removed two of the measuring tapes, but the rover was still overweight. I believe this snapchat succinctly summed up the night
You will get checked by the TSA for having so many tools. Strangely, the TSA placed a very specific tool length—a 6.5in screwdriver is ok but not a 7.1in. Also, you must carry on the batteriesAfter fixing the rover and making some last minute alterations to the code, we packed everything up and left for Nevada.
Arriving in the Black Rock Desert, it was quickly apparent that there were not many obstacles to cross.
The drive from the hotel to the desert is about 2 hours. Before you leave, grab some lunch meat and bread — you’ll be in the desert for the whole day.The only ruts on the desert were caused by cars driving. On the outskirts of the desert, where rockets land, there are no ruts. The next several hours were spent debugging and soldering components to the rover. We did not have a table and there was a sandstorm outside, so everything took place in the small confines of the minivan. Not optimal.
Around 1pm the rover was ready, so we gathered everything up and checked in. This requires weighing the rover and parachute, then proving you can fit inside the payload tube. The rover cannot be any amount over 1050g for the competition, however, you can ask the assigned rocketeer if being overweight is ok. Our rover was overweight, but we asked, and our great rocketeer Dave agreed to let us fly.
It is extremely important to never lose site of the rocket. We designated a spotter who watched as we guided him to the car. Once in the car, it’s a race to drive to the rover which could lead you several hundred yards away from the launch site.
Now you wait for your code to do its’ thing and hopefully drive to the target. Our rover, unfortunately, did not make it.
Lessons Learned:
- Very dusty
- Bring a jacket
- Bring 2 backups of everything
- Soldering in the car is likely a necessary evil
- Make electronics easy to remove
- Make rover faster
- Work much longer on navigation code
- Bright orange parachute worked flawlessly
Created: 18 Jan 2017