Week 1 - Prototypes

Shameless sponsor plug, but also it took me like 5 years to get a robot into the office!
But I finally did it!

Wow - somehow a new season has started, a new game revealed, and a week gone by already! Some time ago, I asked a friend for their thoughts on my blogs. I got the feedback that "It's fun! It's like you're taking the reader on a journey." So, lace up your shoes, and let’s take a walk through our build and competition season for the 2025 FRC game - Reefscape!

As always for the first post, below is a link to the game animation - a high level overview of how to play this year's game. This animation is accompanied by a 160 page manual with all the acknowledgements, rules, regulations, and tournament information. (Most of the manual is outside the scope of this blog post. Possibly because I haven't read it all.)

The first day for all new seasons follows a pretty common flow:

• Eat pancakes (Bearcat specific - but I'd encourage others to do the same!)
• Watch the livestream
• Ooh and aah over the new game
• Read the rules
• Break down the game in terms of robot actions and scoring points
• Define your robot strategy and action priorities

This day is all about 'What' not 'How'. We want a robot to do <some action>, but we don't really know or care how it will do so. This day is also about absorbing as much of the new game as we can - what is allowed, what is not allowed - so that when we start defining 'How' we will perform actions, the robot and strategy are legal to play!

Also common to the process is aligning robot goals with team goals and team resources. Is it our goal and do we have the resources to build a multi-joint arm on a turret that’s fully autonomous and can also make a decent cup of coffee? Sadly no. Do we want to build a simple but effective robot that can earn us a spot at the World Championship? You betcha! (Do we also want it to make coffee? Yes, but we'll get to that later.)

To help simplify things, one process we added to our strategy and priority list was explicitly stating that we won't perform an action. By specifying this, we save design, software, and manufacturing resources, as we won't need to prototype, CAD, test and tune, and build certain subsystems. That time can then be spent making the features and functions we are building, even better.

For Reefscape, our focus will be on scoring enough points to try to win matches, maximizing opportunities in Autonomous mode, and contributing to the bonus Ranking points. On kick-off, we built a priority list as follows:

• Score Coral on Level 4 (with ability to intake more Coral)
• 2-piece Autonomous at our Week 1 event
• Clear Algae from the Reefscape
• Score Coral in Levels, 3, 2, and 1
• Shallow Climb
• Control and score Algae in the Processor

We also gave ourselves flexibility that, if resources provided a "bolt-on" solution for Deep climb, we would alter our priority list to accommodate that. Fortunately, resources obliged, and several Robot-In-3-Days and Open Alliance blogs described simple and effective methods to perform a Deep Climb as Build Season week 1 progressed. We haven't written a new priority list yet, but we will soon.

All that was just Saturday!

Lots of busy hands during the first week.

Our next team meeting was Tuesday, but for us mentors, as I'm sure was true for many students, our brains didn't just turn off after kick-off. (The skill of being able to turn off as soon as you leave the shop, called compartmentalization, does come in handy later in build season.) Kayla and I have been through a few build seasons, seen quite a few robots, and around this time, so much exciting problem solving starts to happen in our heads. 

Out on a single aternoon walk through the neighborhood, robot concepts, space claims, motor counts, actuated movements turned and tumbled until we found a perfect idea. That night while washing dishes we suddenly realized our perfect idea had a flaw, and our brains began to turn and tumble once again. Thinking only about scoring Coral, autonomous, and a climb (our top priorities), we conceptualized <many> configurations and iterations, each building on the previous in search of perfection.

Then comes the hard part - our goal isn't to simply describe the final robot design. We need to walk backwards through all our own chaos and build framing questions. Why did we make this change in our head? What problem did it solve? What new problem might this introduce? As we meet with the team, we use these problem statements to guide design discussion and see what solutions evolve from the group. These questions may involve our resources (Can we manufacture this thing?), may involve design principles (Can we make it simpler? Lighter?), may involve operation (Does this movement interfere with that movement), etc.

Through the week each subteam began specific tasks designed to help us move quickly through build season, and start to supply key information for "How" the robot can perform the prioritized actions we want it to. We have another team goal: finish our version 1 robot by week 4, so we gotta go fast. (We set this goal every year. We'll get there one year, I promise!)

Solid start - chassis with some wires on it!

Software:

New year, new libraries. This year however, we have last year’s robot on a swerve drive base that we can use for initial testing and tuning. (Spoiler: we're doing a swerve drive base again this year.) Over the past year, our software team gained a ton of knowledge and application using libraries for swerve driving, PathPlanner, and how to incorporate vision processing. Part of the team's week 1 schedule was providing a powered chassis to software so we can set configuration values and get our 2025 chassis driving as soon as possible.

Mechanical:

We all have ideas, some good, some less good, and some quite zany. To know if an idea will meet our robot needs, we have to prototype it. This week our mechanical team built prototypes for intaking Coral from the human feeder station, and scoring Coral. In the shop we have a laser cutter, and were really hoping to use it to help make higher fidelity prototypes faster. Mission accomplished!

Electrical:

The main focus for the electrical team this week was preparing the 2025 chassis. We want to get something moving as quickly as possible, so on Kick-off day we made our (slightly trapezoidal) frame size decision. These days, a chassis can get assembled and wired in less than a week. And our team followed suit, delivering a frame with all the core power and signal electronics wired together.

Design:

Design, design, design. Kayla and I commiserated that last year we were in week 2 or 3 before we had a basic image of the robot in our head. This year we arrived at a pretty high fidelity brainCAD on Tuesday. We have many new designers on the team this year, and being able to meaningfully support each of them is proving slightly challenging, however each student has a subsystem to work on. Their jobs include researching existing subsystem examples and ideas, and re-making a suppliers entire CAD model because we're using different tube lengths, and their model doesn't have great variable support. Already, design and software have started talking sensors - including getting a sketch of camera location and view frustum, ensuring we provide a planned home for our camera that can view the details we need it to.

Orange looks so good!

Miscellaneous:

This probably falls under mechanical, but alongside prototypes for robot parts, we need something to practice with to validate our robot parts. This week the team assembled mock field elements: part of the human player Coral feeder station, and started work on a Reef. We also produced a set of trademark Orange bumpers for demo's and other practice. We don't drive the robot on the floor unless it has bumpers on it!

Week 1 Wisdom: Prototyping

(This topic is funny for me, because generally I am very bad at prototyping. So take what you will, or get a good laugh from all these thoughts and recommendations :-D )

For most teams, week 1 is all about prototyping. Generally speaking, prototypes are built to answer the question: Will this idea actually work? The goal is to validate yes or no, and if yes, take that boolean knowledge and put the final system into CAD. Just kidding, but here's the wisdom in prototyping - tests should be simple so you can go fast, but specific so they can be measured. Critically, once something can be measured, write down and store measurements, because those values will find their way into CAD.

This week we utilized - I'm going to claim 3 - prototypes for our Coral scoring mechanism. Prototype 1 was second-hand - we watched Cranberry Alarm's Robot-in-3-Days videos, and they showed that you can build a chute or slide from which to score Coral. So we grabbed a piece of bent metal, some lengths of wood, a piece of PVC, and lots of clamps. Prototype 2 was a rig that we could use to measure angle, height, and horizontal distance of scoring (or failing to score). Our critical measurements for this prototype were based upon the location and angle of the Coral game piece. We brought this information into CAD and built some 2D geometry for roller wheels and contact patches. Prototype 3 was cut on our laser cutter, fitted with bearings, shafts, wheels, and belts, and could be powered by a cordless drill. For this assembly, we kept the angle fixed, and measured/verified height horizontal distance. This time we measured from the roller wheels, as they are an easier fixture in CAD to measure and adjust.

Prototype 2 - measurable in a few dimensions

Prototype 3: Starting to look like a robot part!
And an index finger so blurry that even AI can't get it out!

I am super proud of how we elevated our prototyping game this year - which has already allowed us to move quickly, but we need a few more robot systems before I can definitely say how competitive it is!

It's easy to get carried away, but after week 1 this feels more like 2023 than 2024 for us. And our lead students are keeping their foot on the throttle.

Quote of the Week:

(During a design call, talking about how a Coral ground intake might work.)

Student: "I don’t get it! it just looks like you're intaking a ball!!??"

Every Mentor: <grins>

Kayla and I plotted and schemed during a Maltby Cafe breakfast.
Pro tip: Tabasco bottles help when you need to draw circles.

-B