Week 8
This was a shorter week for us, mostly because on Friday March 3rd, we loaded into Glacier Peak High School for our first competition of the year! We started the week with some final updates/tuning/drive practice, so, we can start there.
This also seems like a great time to really talk about Insight, the 2023 robot from 7627 Bearcat Robotics. Through these posts you've seen a lot of our journey to get here, but let's reflect and dig into the metal on what makes this particular bot tick.
 |
| Slightly bigger carpet than we have at home. |
Unpacked at the practice field, ready to work out the next set of kinks! They came quickly... We use playstation-style controllers for our teleoperated driver and operator to control the functions. While this is a simple robot, there are a few buttons for height pre-sets, intaking and out-taking game pieces, running the pneumatics, and it started as a jumbled mess!
 |
| Full sized field allowed us to practice - including our first time driving on a charge station. |
After a few cycles driving, and many re-mapped buttons, we started to get into a decent groove. Below, we are the robot in the background wearing blue bumpers. The field was shared by a few teams, all of whom we would see at our first event. Watching our preparations and our drivers getting comfortable with the controls, I (and Kayla) felt like we were showing promise - this simple bot could play effectively with some of the more veteran and better equipped teams!
 |
| In the background in blue, running through our program. |
One final checklist item before packing up for the event - sponsor stickers! There's a mantra in Formula 1 - cars that look good, generally perform well.
 |
| Always thank you to the school, students, sponsors, parents, and mentors! |
This robot looks good.
Insight
Let's take a bit more of a dive into this robot, with some necessary context about our team resources this year. Insight consists of three main subsystems; chassis, elevator, and intake. Each was built with specific, measurable goals in mind.
For some fun added context: the team's hours this year were Mon and Weds 3pm-6pm, and Fri, Sat 1pm-6pm. Outside a few sessions at the practice field, we kept these same hours for the entire season. No crunch!
In November, before the season even started, we shared the Citrus Circuits Fall Workshops YouTube videos to our team. Specifically the lessons on Strategic Design, and Simple Robots That Win. If you haven't watched that series, I do recommend you pause reading here and go watch them. When you return, you can critique if and how well we applied some of their teachings!
Chassis:
 |
| The kit chassis, holds everything together. |
In an earlier post I commented that the kit chassis was probably the single item that made our build season as successful as it was. Swerve would afford a team a higher ceiling, but we were starting a lot closer to the floor. Using the kit chassis allowed us to spend a lot more time focusing on mechanisms above the deck, and was easier for our software team to write controls for.
We opted against any cutout in the chassis, or even in the bumper. This game would see full field traversal over flat ground - high opportunity for high speed collisions. We wanted to make sure our bumpers were secure, and as best we could, remove the opportunity for damage to our robot. Intaking over the bumper does add a slight difficulty, but even that gives us a benefit - when we're in the low and stowed position, we can't draw any penalties while driving across the field! (So long as we stay out of the opposing load station...)
Speaking of control, we did give our driver a few buttons to play with. Our basic arcade style drive operated at about half speed, and the bumpers on the controller could either give the driver turbo mode! or slow mode. This would allow for fast cross field traversal, and precise alignment for scoring.
Elevator:
 |
| Lift-y. |
Repeating myself, I know - we bought the 2-stage ThriftyBot elevator kit. We performed as little modification as possible, making it easy on ourselves to follow the included instructions for assembly. The lift is slanted at 60 degrees, and mated via gussets to additional tubes creating a quite rigid a-frame. All of the 2"x1" and 1"x1" tubes are 1/16th wall thickness for low weight. 1/8" wall would have been stronger, especially where extra-loaded bearings wore against the edges, but we did not have the time or machining resources to introduce lightening patterns in the material.
The angle was chosen as one solution to the question of keeping weight low-down, while being able to quickly extend high enough to score on all levels and pick up from the shelf. We optimized for scoring on the high level - we could drive our bumpers into the hybrid node separators, and the forward extension lined up right on the pegs. (We had to drive more carefully to pick up from the shelf and score on the middle row.)
The lift was powered by 2 NEO motors, geared at 9:1, sharing the output shaft. Those motor plates were... fun... to work on. The gearing was chosen again for controllability - the theoretical max motor output would allow full elevator travel in ~0.9 seconds with a 2X safety margin for power, based on the weight estimate of our intake. Software was able to start with low motor powers, and as we improved control, could ramp up the speed. Eventually we hand-wrote setpoint logic with a linear interpolation ramp down, and pid controller to hold the set position.
Intake:
 |
| Spinney. |
The majority of our design time was spent on the intake. As this is the component that interacts with the game elements, that seems to make sense. (At least it does to me!) This final design was influenced by previous years robots, Open Alliance build threads, the EveryBot, and one of the FRC golden rules: positive acquisition. The intake consists of four identical polycarbonate plates, 2 per side that protect the gears and belts, and the gap between where the game elements will be collected. For the cone, we used VersaRollers and 3" compliant wheels, the cube uses 2" compliant wheels on a hex shaft and a fixed 3/8" shaft. (I don't remember the specific compression off the top of my head...)
While prototyping roller intakes during week 1, we accidentally found out that we did not need to actively rotate both rollers to hold a game piece. (The uh, second drill fell off the shaft it was driving, yet we still acquired the cone... Can we use this knowledge? Yes!) Only 2 of the 4 game element contact pieces (the spinn-y bits) are driven by the NEO motor - the 3" wheels for the cone, and the 2" wheels for the cube. From a software and operator standpoint, this allowed wonderful simplicity! Only 2 transmissions belts were needed, and both rollers spin the same direction - the operator does not have to think about which game piece they are acquiring, they press the same button! In between the 2 driven shafts is a free-spinning roller covered in grip tape that helps hold the squished cone. The downside to this idea is to make space for an additional roller, we had to elongate the entire intake. Weight up high is the enemy, but the afforded simplicity was worth the tradeoff. (Just drive slower, we said to a teenager....)
 |
| Intake CAD |
The powered shafts were geared at 3:1 then 2:1, ensuring that the wheels surface speeds were greater than our driving speed. This meant we could pick up game elements while driving forward at them. Our cone wheels were 3" to account for the drop-center wheel chassis tip. Regardless if we were tipped forward or back, we would make contact with the cone and intake it. The intake was deployed and stowed with pneumatics - first time using them for Bearcat Robotics, and generally a solid choice for 2-position articulation.
Now that we're all familiar with this robot, let's bring it to competition already! Our first event was Week 1 - Glacier Peak! We had some fun. :-)
No comments:
Post a Comment