CavBots 7492 - 2024 Open Alliance Build Blog

[CavBots7492]

Pre Season Post

Hey everyone! FRC 7492 CavBots is pleased to announce that we will continue to be a part of the openalliance for the 2024 - Crescendo season, and are excited to be a part of it for years to come!

As per tradition, a little backstory on us: We were started in 2019 in The Woodlands, Texas, a suburb just north of Houston. This will be our third year participating in the Open Alliance and we hope to maintain a great work ethic in regards to posting frequently, descriptively, and as truthfully as we currently envision. To continue the blog’s integrity, we will show The Good, the Bad, and the Ugly (!!) that we will inevitably endure throughout this season so that we are able to teach and guide others as well as show our families and sponsors exactly what we’ve gone through this season.

Good luck to all the teams competing this upcoming season!

We will be competing at:

• Katy (Week 1)
• Fort Worth (Week 3)
• Hopefully FiT District Champs and World Champs!

Resources

We will keep these as regularly updated as possible. Notes: We keep our CAD files in Google Drive. There will be uploads of pictures and videos in case you’d like to see them! GitHub is our public code repository so as the season progresses our code will become fleshed out and more complete.

CAD -

https://drive.google.com/drive/folders/1TxDoyBnc40ir4kZS2P8JcmPE9mHHvKbn?usp=drive_link

GitHub - CAVALIER-ROBOTICS · GitHub

Website - https://www.cavbots7492.com/

Written by Sam (Co-Captain), James (Co-Captain), and Daniella (Build Lead)

[Connor-8177]

Vector will also be at the Katy event, looking forward to competing with you guys again.

[CavBots7492]

Post Kickoff (1/8/2024)

KICKOFF AND DAY 1
We hope everyone had a great kickoff this week! Sorry for the late post, we are all very sleep deprived… In terms of us, though, we were able to have a great discussion and came up with a few good ideas. For the most part, we believe we were able to land with a relatively good design this year, and hope that we can execute with careful outlining and planning.

DESIGNS THAT WE CAME UP WITH

These are designs that were brought up by members of the team but we decided to scrap as of right now, maybe another team would like the ideas brought up by our group.

• Slingshot style shooter
• Piston/Spring powered shooter
• Intake notes under the bumpers through the middle of the bot

CURRENT DESIGN PLANS

Right now, the plan is to have an intake that retracts back into the bot to feed into the shooter, that will then be able to pivot and shoot out into the speaker and the amp. The intake will contain sensors for detecting notes, and retract to load the shooter. The plan is for it to be able to have about 130 degrees of rotation in order for it to be able to shoot notes into the speaker and push them out to the amp. For our climb, we plan on using two telescoping tubes on a pivot to align the robot to be easier to climb with other bots. At this point, we believe it to be in our best interests to not go for the trap but we may revisit it later.

For Vision this year we are using 2 Limelight 3s, one forward facing and one backwards as well as a camera for the driver due to the stages being so in the way.

PLANS FOR THE NEXT FEW DAYS!
Our next few meetings will be looking into the best wheels and compression to use for our intake and shooter, as well as finalizing design ideas and preliminary 3D CAD. We hope to reach the end of Saturday with a near complete, if not finalized CAD, in order to pump out machining all of the following week.

PROTOTYPING
We did some intake prototyping today and discovered that ⅛” compression is best for intaking notes without the risk of tearing them up and having them get stuck in the intake.

CavBots 7492 2024 intake prototype testing 1

Here is our build schedule for the season (Our competitions are in Green).

Thank you for reading! We hope you guys had just as much fun a time planning out for this new game. Feel free to ask any questions, comments, or concerns you have in mind.

Written by Sam (Co-Captain), James (Co-Captain), Daniella (Build Lead)

[CavBots7492]

Day 14 (1/20/2024

We haven’t posted in a while because we’ve had some delays due to weather and school cancelation but we’re back to work now and we’ve made a lot of progress.

MECHANICAL
Our CADders have been hard at work the past week and a half and our CAD is almost complete. All that’s left to do now is the fine details.

INTAKE
For our intake we will be using versa rollers with surgical tubes stretched across for the initial pickup and then timing belts direct the note into the shooter. To help this, we will add vectored mecanum wheels to better push the note towards the center.

SHOOTER
For our shooter we are doing vertical flywheels with separate NEOs spinning the top and bottom separately. These will be run with a near 3:1 reduction, as not much reduction seems to be needed. We also have a NEO550 to direct the note from the bottom of our shooter into the flywheels. For our flywheels we are using 3” stealth wheels.

CLIMB
For our climb we will be using 2 stage telescoping tubes, very similar to what we did in 2022.

MACHINING/ASSEMBLY
We are currently in the process of machining out all of our parts. So far, we have cut out a few PC sides and aluminum gussets, all of which we’ve had no problem with. Thankfully, although the weather set us back, we should be on schedule to finish machining all of our parts by the end of week 3 and begin electrical wiring soon.

PROGRAMMING
This week our programming began to get limelight field pos setup and starting to tune our speaker aiming values. We are currently waiting on a google coral to come in to start working on automatic note detection and aiming so as to help our drivers/auton.

Here’s a screenshot of our full CAD we hope to publish it within the next week when everything is finished.

Thanks so much for reading our post and feel free to comment any thoughts you have about the robot.

Written by Sam (Co-Captain), James (Co-Captain), Daniella (Build Lead)

[FredK]

Have you noticed any damage to the notes from the timing belts?

[xcer85]

We did notice some above normal wear on the notes but nothing too major that it would damage the piece

[Waltja]

Hey y’all I’m James, one of the co-captains of 7492 Cavbots and we’ve seen a couple CD posts about how to find shooting angle and we wanted to share our solution. The Java example code below is meant to be used as is and needs no modification unless your shooter angle is adjusted by an elevator or other such mechanism.

One of our programming members has worked since nearly kickoff on a collection of equations that we calculate locally on the Rio, but will eventually be done on a Pi 4B, that gives us the ability to not only constantly calculate both our robot rotation and shooter angle, but also takes into account our robots velocity and properly offsets for it.

This isn’t our full code, we will show that off in our CAD/Code release, but this is a quick mock up of the basis of our equations. This Example Code takes in your robot position in field coordinates and your shooter speed (in m/s) and gives you a resulting angle to the speaker and the required angle of your shooter.

Thanks so much for reading, if you have any question about implementing or having other issues we would love to work with you to resolve the issues.

Written by James, Equations derived by Hayden (Programming Junior Lead)

[Waltja]

Hey y’all, it’s James again and unfortunately this time we have to talk about our season. We have all been pretty busy trying to get this bot out and don’t have the resources (mentors or business members) to either help manage things or to help write OA posts. As of now I plan on stepping back a bit and filling into a more management/mentor role which will allow me some time to write posts and keep up with questions anyone may have for us. Little tangent, shoutout to 2582 for reading our posts, however intermittently they are, and showing me someone wants to read this. Getting this back on track, after our first day at Katy most of our leadership and I made the decision to do some pretty major redesigns.

To get to the meat of it our current design will not be competitive at state, and hopefully worlds. For those of y’all who saw us at Katy we were kinda all over the place, having major vision issues and losing the ability to use our shooting equations which we had planned on carrying us pretty far. This required us to hard code in a setpoint for subwoofer shooting mid competition and throw out some other functionality. We got notes stuck in us multiple times, not because someone missed a shot but just because of basic subsystem malfunctions, something that should never happen. In our new design we will change our intake to a direct feed to the shooter, while keeping our back loaded design. This allows for the complete elimination of the previous issues. Our climbers worked perfectly and honestly with everything else we are changing “If it ain’t broke, don’t fix it” fits this perfectly. We are going to add limit switches so that we can run them a bit faster without fear of breaking the string we are using. We have some electrical changes thanks to 8177 Vector insight that we will be implementing. This year we are absolutely blessed with a team of amazing programmers, one of which has been working on a Jetson Orin Nano running VSLAM to replace our current limelight 3s and the plethora of April tag issues FIRST is having right now. This includes some cool Intel cameras that I will talk about a little later when we get them working. Unfortunately we won’t have any of the major mechanical revisions done before Fort Worth, but we do hope to have our Jetson up and running as well as some minor fixes to push us across the line to qualify for State.

Sorry for the block of an explanation but I wanted to lay out all the issues and changes that we saw and want to implement to give other teams a look into what a process like this may look like. Our team has a wonderful base of sponsors and parents that allow our students to undertake a change as big as this mid-season. To be very clear, this is not something that I would recommend doing to any team that bot currently works, and while I do realize we fall into that category, our team believes that to accomplish our goals (qualify for worlds) something needs to change. If you have any more questions about our reasons and choices that we believe make this change necessary, please message me or one of our wonderful team on CD or on our discord and we’ll try our very best to help explain in more detail.

Thank you so much for reading my post, I plan to make the next one a bit more upbeat with the inclusion of some designs and hopefully programming updates.

[Waltja]

Hey y’all it’s James, and we have some great stuff to share. First off a little explanation of the state of our redesigns, the current plan is to finish the design before Fort Worth. At Fort Worth, if our current bot performs as well or better than expected we won’t build these revisions for state/worlds since it wouldn’t be worth it to sacrifice more time/money on a new bot when this one works. If somehow we just barely scrape into state and don’t think we have a chance for worlds we plan on churning out this new bot for state to help qualify for worlds. But enough of the planning let’s get to the fun stuff

Upgrades:

• Uses a “direct feed” intake that loads the shooter right from the ground
• Adds 2 static rollers between the shooter and intake to help guide note
• Compresses note horizontally from 1 in in the intake to ½ in in the shooter
• What I call a Spectrum frame, yes 3847 Spectrum, using raised 1x1 tubing to help support the static intake rollers and the intake gussets
• Raises the shooter pivot tubes, but not the pivot itself, to allow electrical to work completely in the belly pan
• Climb now includes a hard stop/limit switch to avoid breaking any more string cough cough

INTAKE

• Uses 3 powered shafts and 2 free spinning to lead notes into our static rollers
• Has an L-shaped cutout in the intake to allow notes to be “spun” onto and not smacked away if your driver misses

SHOOTER

• Adds top/bottom kicker wheels to control notes better
• Changed the shooter wheels to increase grip on notes as we noticed some slippage on our current ones
• Uses an A-Frame design for better stability

Besides this, our other 2 designers have been working to fix up Sentinel and give it a couple of touch-ups for Fort Worth. Some quick intake changes and electrical moves but the biggest change is the addition of an Amp Bar that will hopefully prevent what happened at Katy, that is the field Amps being different so our tuned Amp shooting didn’t work. This is the last thing we need, in our opinion, to become a good first pick for the top 4 alliances, something we need if we plan on getting to worlds safely.

One last thing before I finish up this post, our Programming team has been hard at work, even until 11 pm on Sunday, to get some cool vision stuff up. As promised in my last post, it’s time to explain. For our vision we are using a Jetson Orin Nano to process all of our camera feeds including 1 Intel RealSense D435i for VSLAM, I’ll explain that in a second, 1 RealSense D455 for note/bot (yes like other bots on the field) detection, and more than likely 1 USB or RealSense T261 Camera for more VLSAM/AprilTag detection. Currently, we only have the D435i but our D455 should be in by Tuesday and running by Fort Worth.

VLSAM

For those of y’all that don’t know what this is VSLAM stands for Visual simultaneous localization and mapping, fancy words for it looks around and maps the field while also detecting its change in position. Currently, we are using Isaac ROS on the Jetson and pushing the localization data to network tables via SSH. Honestly, this project has been crazy, we all assumed it would be a 1-2+ month project with tons of complications but somehow it just worked on the first try and has been working ever since. We had minor issues with pushing the data to the RIO but after some finagling we got it working. If your team wants to try this my advice is

DO NOT

At least not right now, let me explain. This is something that probably only worked because of the abilities of our team. We have someone incredible at vision/Linux/ubuntu programming, a master of Desmos who I’m pretty sure has made more graphs than met people, and a WPILib wizard. These people made this possible in just under a week and are still working to iron everything out. Once we finish or after worlds, whichever comes first, we will work to release a little guide for teams that want to move away from Limelight’s “black boxes” and have more control over their issues. If for some reason you still think trying this right now is worth it, it’s not, we are more than happy to help and impart the mistakes and solutions we found.

Well, that was a bit longer than I would have liked. I may have forgotten to post last Thursday/Friday, so this is kinda 2 posts in one. As always if you have any questions please message me or one of our wonderful team on this CD thread, privately or in our Discord and we will try our best to answer. Thank you so much for reading and good luck to all the teams competing with us in week 3.