Recycle Rush Reflections on Design

Hello readers! My name is Tanay Trivedi, a rising college freshmen at the Cooper Union for the Advancement of Science and Technology out of Bridgewater Raritan High School. I am affiliated with my former FRC Team 303: The TEST Team. This article is a quick recap and discussion of the 2015 FRC Game Recycle Rush, and how all teams, powerhouse or not, should use this experience to further their performance in the future.

In most seasons of FRC, teams use the ending of the Indiana Robotics Invitational as an end of season review, and ask questions like “What did we implement that was new this year?”, “What things worked, what things didn’t?” and “How should we experiment in the off season to aid in our performance next season?” However, this year has thrown the FRC universe off tilt. No defense meant light and under-powered drive trains; for the first time in four years I have seen two-CIM tank drive trains on high performance teams. Driving dynamics changed as well. Teams preferred slow and careful driving compared to the high speed aggression seen in 2014, invited by an open field and increased motor caps. So how do teams proceed this offseason?

1) Do not abandon ambitious drive train projects just because this year had little to no driving involved. As Karthik would say, the highest priority of an FRC Team in the build season is the drive train, simply because moving is the first step to scoring. Drive trains should always be evolving, if only in the software, because small increases in efficiency in moving from Point A to Point B yield large results in scoring points. The first exhibit of my admittedly anecdotal evidence is my own team. After the disappointing performance of HTD belts in the 2014 season (by disappointing I mean ripping each other apart on double pulleys, pulling wheel blocks in every match, etc etc), we decided to do an all gear drive inspired Team 341 and Team 25. This yielded incredible results at the fall offseason event Brunswick Eruption, as our drive train was more efficient, never failed, and as a byproduct, bought our team useful machining and CAD practice in the offseason. How did this help us in the 2015 season? Well, its efficiency boost was utterly useless to us.

With our four-CIM, single speed 10fps gearboxes we were hardly pushing the edges of physics and didn’t require the extra efficiency to help us attain top speed. HOWEVER, it was not all for naught. We had zero drive train failures this past season, which is more than we can say for any of the other 16 seasons 303 has participated in FRC. Having one less thing to worry about during an FRC season is a success to say the least.

2) SOFTWARE SOFTWARE SOFTWARE. Making a drive train more efficient and more powerful mechanically seems to be the end goal of all offseason projects nowadays. However, what most high school team members do not understand, or ignore due to difficulty, is that efficiency and precision can be increased dramatically by implementing “funky” drive code, especially when maneuvering. Any person watching 254 pretty much anytime in 2014 can attest to this. Just watch this video of 254 in the semifinals at Waterloo, and you can see them rock the curved path between their alliance partner loading the ball onto the field and their passing position right next to the truss, and then right into the fray, disrupting the opponents cycle.

If you want detailed information regarding their control software, check out their presentation at champs, which is posted here.

Now, the latest and greatest in drivetrain software development this year in FRC was that of slide drive, or H drive. It involves having four omniwheels in the corners of the frame facing in one direction, and an omniwheel in the center of the frame oriented perpendicularly to the rest. Most people are familiar with drive system, however they are not familiar with how complex it can become to strafe perfectly to the right and left. To keep the physics simple, when the center of mass is fore or aft of the point of contact on the H wheel, the robot will rotate while strafing. To correct for this, one can simply run the normal omniwheels counter to the rotation at a predetermined speed, thus eliminating the rotation. However, if the driver has strafing speed scaled and determined by a joystick, the rotation value changes, thus requiring different amounts of countering. Team 303 encountered this problem, and utilized the new MXP port sensor called the navX, and its multiple gyros to solve it. A detailed discussion on H drive is found here, but the important takeaway is that offseason experimentation with things like H drive and swerve, or even just more complex feedback control for tank drive systems like 254’s will eliminate experimentation time during build season. If you have any doubts about the need for precision and maneuverability in a game like Recycle Rush, check out all the dropped stacks at IRI. Nuff said.

3) Use the offseason to convert to new technologies and systems. From extruded to sheet metal. PWM to CAN. Thinking of robots in your head to actually using CAD. Java to LabVIEW to C and back again. FRC has a constantly evolving technological landscape, with opportunity costs at every step. However, it is these types of engineering decisions that students are being trained to cope with, and it is exactly the kind of decision that needs to be made in the offseason. For example, lets run through the third case of anecdotal evidence from my alma mater, Team 303. In the offseason between Aerial Assist and Recycle Rush, while we were designing the “All Gear Drive” for Brunswick Eruption, we realized that we needed to use sheet metal to be accurate with our gear spacing. Thus, we hunted around until we found a laser sheet metal manufacturing company that was willing to take the CAD model for the drivebase and cut and bend the aluminum into our specifications in the first two weeks. Plenty of FRC teams do this, but we had only realized what a gold mine we had landed ourselves in. We no longer had to use our own team resources to build the drivebase from scratch during the critical prototype and manipulator construction phase, and thanks to our comprehensive CAD model we only had to assemble.

Now, this is not to say that sheet metal manufacturing is the “end-all-be-all” of construction methods, because plenty of things can go wrong. If your manufacturer has specifications on their bending tools that you don’t understand or take into account, you can run your build season into the ground and only have a few days to recover. The only way to eliminate such possibilities is to have a mock build session, and this is where the offseason comes in. If you can engage in an offseason build like our team did, which included a full rebuild of a robot to compete in an offseason competition, you would be in the ideal scenario, but at the very least take a single component, like the drivebase, or manipulator, and completely revamp it. This will give you not only extra manufacturing practice, but also allow you to optimize your build schedule when it comes to build season.

Teams should be utilizing their offseason to find new manufacturing techniques, write new software algorithms to optimize driving maneuverability or experiment with different kinds of drivetrains. All of these experiences at the very least will enhance a student’s experience level with engineering, and at the very most actually cut short the experimenting stage of build season.