FRC 4255 The Robodores Build Thread

Hello! It’s exciting to announce that Team 4255, The Robodores, will be participating in the #openalliance in 2022! We’ve been interested in an open build for a while, both for the possibility that we might provide some utility to other teams in a year where many are in a state of rebuilding, but also that organizing, sharing, and getting feedback on our process could be really helpful for us.

The Robodores work out of the Monterey High School robotics lab in Monterey CA. We are lucky to have access to a Velox CNC router, two 3D printers, and basic shop tools (bandsaw, drill press, belt sander, hand drills, etc), as well as a sponsorship with a local metal shop to get parts cut on a waterjet. We are a relatively small team of about ~10-15 students with 3 mentors. The team is very student run, with mentor/parent volunteer support generally on the non-technical side. The Robodores did not participate in the 2021 At Home Challenges, instead building an FTC robot and competing in Ultimate Goal. We are currently feeling the effects of a gap in knowledge and skills between seniors who experienced the pre-COVID 2019 season and newer students, with 2022 being their first official season.

During the 2021 off-season, we aimed to address this knowledge gap and shake off any cobwebs by building a completely new bot for Infinite Recharge. Despite some late event cancelations, we were able to attend Madtown Throwdown with two drive teams, and overall had a great experience. An album documenting the highlights of this fall can be viewed here.

I will be documenting our 2022 season at least weekly, likely more often in the first few weeks, in this thread, as well as posting CAD/Code when available. We will be attending our local Monterey Bay Regional (Mar 23-26) and the Silicon Valley Regional (Apr 7-10).

Good luck to all in 2022, more updates to come.

11 Likes

For the first time in 2 years, today our team tuned in to the kickoff stream live and in-person from our shop!
Before that though, we opened a late Christmas gift or two:

Super excited to run the new PDH, pneumatics hub, and radio power module on our 2022 bot.

Kickoff:

Although I got a few predictions right (new balls of different size/texture, multi-bar hang), there were definitely some surprises with in the Rapid React Reveal. Another ball indexing/shooting game feeling a little tired, but many well developed concepts can hopefully be applied.

Game Analysis:

We took a simple approach to strategy similar to past years, making and debating the relative frequencies and difficulty of implementation for every method of scoring points. Weighting total point scoring potentials versus difficulty gave us a rough idea of value, shown below:

We compared 3 potential robot archetypes: climbing to the transversal run and scoring in the lower hub, and climbing only to the mid or low and scoring in the upper hub

Some Takeaways:

  • Climbing past the mid-rung seems extremely difficult for only 9 extra pts (cut to only 3 if an extra cycle can be completed in endgame as a result)
  • Lack of human player loading beyond the first ball seems extremely strange, basically mandates a ground intake and therefore eliminates a major advantage of lower hub scoring (same height as loading station)
  • Prototyping will be needed to get a better idea of the difficulty of upper hub shots

Robot Goals:

Based on the analysis above, we decided our robot must do the following:

  • Shoot into the Upper Hub from 1-2 positions reliably, likely without turret/adjustable hood to simplify design, programming, and strategy
  • Intake from the ground, basically the only viable option this year
  • Almost definitely use a West Coast style Drivetrain similar to previous seasons, if all else fails we want to ensure we have one of the strongest drivetrains in any given competition
  • Be done quickly (<6 weeks), test, drive, break, and modify as soon as humanly possible

Design Brainstorming:

  • Considered various shooting positions (protected zones, against fenders, against walls)
  • Ideas for intaking/indexing include 2019 Everybot inspired bumper gap, possibly improved with vectored intake wheels (inspired by 2910 2020/21), over the bumper intake and “v”-shaped rollers (4414 2020/21), or not indexing at all (2132 2021).
  • Ordered WCP GreyT Telescope immediately after the game reveal, would be more than sufficient for mid-rung climb
  • In case Telescope doesn’t arrive in time, we are considering a pneumatically or tension actuated arm with detachable hook on winch (similar to the first stage of 1678 2020 Mk 2 climber)

Finally, we decided that we’d prioritize prototypes for both over-bumper and vectoring intakes, “v” hopper, and a hooded flywheel shooter.

Future posts probably won’t have quite as much detail, as this took a while. Good luck to all teams, and let us know if you have any questions or feedback.

2 Likes

Day Two:

Today was focused around the construction of our hooded flywheel shooter prototype. Use these links to view CAD and photos/videos from testing.

Design:

  • Geometry allows compression to be adjusted through slots
  • Accommodates wheel diameters from 4" - 8"
  • Polycarbonate side panels cut on router, mounted to 2x3 frame
  • Up to 2 Falcon 500 motors belted to the flywheel at 1:1
  • Simple control system that can be used on future prototypes

Testing:

  • Without a full Upper Hub mockup, we used a basketball hoop as a reference for height
  • With 1 Falcon, 4 in Colson wheel, consistency was decent, hitting the rim multiple times in a row, went in once (unfortunately no video)
  • Compression changes from 0.5 - 1.5" increased distance slightly
  • Switched to 8 in pneumatic wheel, added second Falcon, distance/height seemed to increase but vibrations prevented consistent testing conditions
3 Likes

Week 2 Update:

Sorry for the long gap between updates. I’ll try to summarize here our progress over the last week and a half.

Prototypes:

Shooter:

Our programming subteam was tasked with building a quick mockup of the upper hub for shooter testing, which went surprisingly well despite some rigidity issues:

Unfortunately, initial testing on Friday and Saturday was extremely inconsistent from the fender position, fluctuating between hitting the rim, undershooting, and going in.
We eventually settled on an angle of 3-3.5 degrees, with about 3/8" of compression, at nearly max rpm (6000).
On Monday, we tried tuning a velocity PID controller in an attempt to improve consistency, but even filtering and gradually ramping up the target, oscillations wouldn’t decrease. We also experimented with a modified bang-bang controller, but it required constant retuning for changes in target velocity and battery voltage. This Friday we will try implementing a constant feed-forward term to our PID controller to hopefully make it easier to tune. If anyone has had success with this we’d really appreciate any advice.

V-Hopper:

Over last weekend, I spent some time designing a prototype for a serializer/indexer based around powered rollers in a “V” shape, and a powered floor, using 1.25" polycarbonate tubing (similar to Vex VersaRollers. CAD is posted here (prototype branch), some details have since changed.
On Tuesday, we CNC’d out some polycarbonate plates, and by Thursday it was fully assembled and ready to test.
Results of testing showed that these balls, unlike Power Cells, had very little grip against bare polycarbonate rollers. Additionally, deadzones existed on the floor when balls centered on the plates between the inner and outer rollers, even when feeding a single ball.

We made several changes to the setup/rollers:

  • Holding a plate compressing the balls from above, with negligible improvement.
  • On a slight incline, much more effective but would be more complex to package.
  • Added strips of grip-tape to the center rollers to improve contact radius/traction, which worked much better at pulling the balls in and centering two balls at once. Some deadness still existed near the center.
  • Switched out the “V” wall rollers for compliant wheels, which we found were much grippier and had greater contact area, and wrapped the floor rollers with electrical tape. We also experimented with running one side faster than another, which seemed to be much more effective at centering two balls.

We’ll probably be wrapping our floor rollers in surgical tubing rather than electrical take on the final intake, but other than that the geometry will remain mostly unchanged.

Robot CAD

By this point we have a mostly finalized version of every mechanism, assembly available here.

Mechanism summaries:

Drivetrain

  • CAD
  • 6 wheel WCD
  • 1/8" center drop
  • 3 Falcon gearbox at 6.40:1
  • Theoretical 17.4 ft/s free speed
V-Hopper

  • CAD
  • Powered by single Falcon
  • 1.25" floor rollers (VersaRollers), wrapped in surgical tubing
  • Floor/walls linked by bevel gears
  • Wall rollers probably to be replaced by 2" compliant wheels
Shooter/Tower

  • CAD
  • 3 degree fixed angle (Fender shot)
  • 2 2"x4" WCP flywheels
  • 2" accelerator/kicker wheel
  • 2 Falcons at 1:1 belt ratio
  • Dual HTD timing belt conveyor powered by single Falcon
Intake

Climber

  • CAD
  • Single WCP GreyT Telescope
  • 2 stage variant
  • Single falcon at 15.34:1
  • Secured to bottom of chassis frame

Fabrication

Chassis
  • CNC’d frame tube, gussets, gearbox plates
  • Frame assembled Tuesday
  • First year using hole pattern tube

It the coming week we will be focusing on fabricating parts for final mechanisms, wiring/testing the chassis, and ordering needed COTS parts.

9 Likes