5013 the Trobots 2023 Charged UP Open Alliance build thread

[Bmongar]

Welcome to the Trobots 2023 Charged Up build thread.

Why did we decide to join the #OpenAlliance?

• Commitment device to produce better documentation and keep our CAD up to date.
• Peer review. Sometimes we have bad ideas, the faster we realize it the better.
• Visibility. We want to get our name out and get known better outside of our region.

Team Links

• Home Page - Trobots.org
• The Blue Alliance
• Github - Code Repository
• OnShape CAD Team
• Our YouTube Channel
• [Discord]Open Alliance Discord Channel
• Team Donation Link - In case you find something useful and want to tip us.

2023 Competitions

• Heartland Regional (KSLA) in Olathe, KS Week 3 3/15-218
• Green Country Regional (OKTU) in Tulsa, OK Week 3 4/5 - 4/8
• Hopefully FRC Championship

Since our last update in our initial announcement thread we have had a parents meeting where we set expectations. We also had interest from 3 adults to become new mentors.

Programming
Still researching Photon Vision. We have found their 2023 Beta thread useful . We are leaning towards a Beelink mini PC as a coprocessor.
We have also been following the Pathplanner updates. We used it last year and it looks even better this year.

Mechanical
We have ordered 5 new conversion kits to turn the SDS MK4s we have into MK4i
Mechanical has been considering the new COTS elevator options.

Electrical
Our overall focus for the 2023 season will be reliability. To that end we will be doing a lot more soldering and using of conduit.

Strategy
We noticed that our friends, neighbors, and fellow Open Alliance team 2357 System Meltdown were attending the same regionals as us. So we decided to from a scouting alliance. Details about that are still being worked out.

Build Season Meeting Schedule
Monday - Short meeting after school recap last week set focus for coming week
Tuesday - After School until 8:00 dinner provided build meeting
Thursday - After School until 8:00 dinner provided build meeting
Friday - Short meeting after school to set up for Saturday
Saturday - 9:00a - 5:00p lunch provided. Most of the work gets done.

When we get closer to kickoff we will be releasing a kickoff schedule.

[diswed0226]

Can’t wait to watch your team’s progress through the build season! It’s been crazy how much your team improves every year. Could you elaborate on your plan for using conduit for electrical? Are you going be routing wires through the metal conduit you can buy at home Depot?

[Bmongar]

@diswed0226

Are you going be routing wires through the metal conduit you can buy at home Depot?

Nothing as drastic as that.
I was referring to split loom conduit which will either be mounted by yet to be devised 3D printed mountings or with judicious zip ties and adhesive zip tie anchors.

[Bmongar]

Pre Kickoff Update
The plan for 1/6

• 10:30 setup for lunch grilled cheese
• 11:00 Watch kickoff show / Lunch
• 11:45 Game reveal
• 12:00 Read rules!!!
• 1:00 Small group game analysis
• 2:30 Rolling chair/human game simulation
• 3:00 Large group game analysis
• 4:00 necessary robot functions
• 5:00 Go home, think about the game, research mechanisms.
  • Unofficial FRC Mechanism Library
  • Spectrum Mechanism Library
  • Read Chief Delphi

Sunday
1:00 pm Return to school

• New perspectives on game strategy
• Robot functions
• Refine game strategy
• Propose mechanical/computational solutions
• Research on Chief Delphi and Spectrum Mechanism library

Season goals

• Develop student skills and confidence
• Win a team attribute award
• Build competitive robot that ranks top 3 at regional (focus on durability)
• Win a regional championship or wildcard
• Get Picked for divisional playoffs at worlds

Weekly schedule
Monday 3:00-5:00 planning
Tuesday 3:00-8 :00 build meeting with dinner
Wednesday No Meeting
Thursday 3:00-8:00 build with dinner
Friday 3:00-4:00 setup
Saturday 10:30-5:00
Sunday No Meeting

[Bmongar]

Very exciting game this year.

We did a little experimentation to see if the drivetrain could support itself passively on a sloped charging station by making a diamond (or x we didn’t have this drivetrain programmed for that). It could be the alignment would need to be pretty good so maybe not reliably. We may potentially need a break of some kind.

Our students ran a couple of rounds of roller chair game simulation.

Roller Chair Charged Up 1

Roller Chair Charged Up 2

Then we had a discussion of capabilities our bot might have. We categorized these as Must, Could, and a stretch category of Would Love to Do. As a team we usually try to accomplish Must and Could

• MUST DO
  • Build on Time
  • drive swerve SDS MK4i
  • pick up both game pieces from double substation
  • place game pieces on nodes
  • Top/Mid
  • drive on to charging station.
  • reliable auto odometry
  • low center of gravely
  • passive defense (protect internal components)
  • read align w/ tags/tape?)
  • effective match strategy
  • touch it, own it
  • solid, easy to repair, reliable
  • partial handoff 01/30/23
• COULD DO
  • pick up game pieces from floor
  • place game pieces (high node)
  • engage on charging station,
  • plant/stabilize on charge station
  • auto engagement (balance)
  • auto place game pieces (in tele)
  • align w/ April tags
  • place/ pickup/engage auto
• WOULD LOVE TO DO
  • pick up game pieces cones all angles -Orient cones
  • April tag localization
  • minimize footprint to 3x engage
  • On the fly trajectories
  • 2xx auto(+)
  • auto pickup stray game pieces during auto

One thing that came our of our discussions was that this year the game is more of a programming challenge than a mechanical challenge.

One of our alumni found a paper on automatic placement and retrieval of traffic cones done by the Colorado Department of transportation.

Our students are now sent home to research how to accomplish these tasks.
We asked them to review previous games specifically

• Logomotion
• Rebound Rumble
• Power Up
• Destination: Deep Space

We will reconvene tomorrow at 1.

And also to check out the resources posted yesterday.

[Bmongar]

We had a shorter meeting today just 1-5. We talked about mechanisms to accomplish the tasks we want to do. Few things were decided and intake is still up in the air.

What we did decide on were the drive base would be a mk4i swerve, rectangular shape, and an aluminum bellypan with electronics mounted on it. We will protect all of our wires with a harness.

The super structure will most likely be an A-Frame. The elevator with a boom seemed too tippy and with too much of a cantilevered load.

We will have a telescoping arm probably a 2 stage Thrifty. 2 stage to allow for more tube overlap and greater rigidity over the distance. We have a few diagrams of the rough dimensions of 1 stage . After reviewing the geometry it was decided that a 2 stage was a better solution. If you have any questions @DChilson or @Embily would be better for an answer.

Our intake discussions are too varied and indefinite to give useful information. Though the discussion was entertaining. We will begin prototyping some of them this week.

The programmers worked on auto path ideas and what concerns there would be @tylurr will be able to cover more of what was discussed if people have questions.

The diagram may not help much, but we are going to work are autos iteratively easy first then out, Move->Move+Place->Move+Place+Engage-> Move+Place+Pickup+Place … We prefer going around the charging station but did decide that going over was a definite possibility but would take some caution.

While we are waiting for our Mk4i conversion kits we will be using our Rapid React bot Sidewinder to test our pathing. We plan on using FRCPathPlanner again this year. PathPlanner currently doesn’t support the fact that field is only mirrored on the x axis and not the y axis thus the red and blue paths aren’t identical as they are in most years but @mjansen4857 said they are aware and working on solutions.

Our GitHub project for ChargedUp is setup but pretty bare at the moment.

The onshape robot assembly has also been setup but also not very exciting at the moment. Onshape

I missed this afternoons strategy discussions but @saatb is our strategy guru and might be able to give an overview over what concerns and questions we have.

We have a folder setup for out build photos and videos . There are different folders for different weeks, Week one has some content.

[jedgar6]

@Bmongar

Out of curiosity are you guys planning of intaking from the player station? (most likely yes) Have you looked at the angles involve? We had a similar design and the angle you pick up from the stations…requires a little configuring (at least us).

[DChilson]

Our preferred pick up point will be the shelf and the sketch shown is able to move to each of the game positions while only extending outside the frame perimeter in a single direction. What additional configuration are you looking at needing to do?

[jedgar6]

Sorry if I was a bit vague and excuse the crude drawing (also imagine the bumper is there).

I know the robot could reach it but do you plan to extend the arm a little to help with a better angle or mount the point of rotation higher? I know there’s a few answers but it’s something that adjusted our design.

[saatb]

Strategy

---

Hi! I’m the strategy lead on team 5013, and I’ll be writing all the strategy related updates for our 2023 Open Alliance build thread.

This post will be broken up into two parts: a (very late) pre-kickoff/offseason summary, and post-kickoff.

Pre-Kickoff/Offseason Recap (click to show)

We went into the off-season with not much to do strategy wise. We had implemented a new scouting system for the 2022 season, and only had some minor things we wanted to tweak before the 2023 season.

I learned that we could export data from our scouting tablets to the laptop via bluetooth without much trouble, which was something that we wanted to accomplish. Our old system had us using USB fobs to transfer data between devices, which was a slow and tedious process considering we were running 8 tablets.

Before Cow Town, we trained the team on our slightly improved scouting system. Our plan was to have all the new members scout ~5 matches at Cow Town so they could get a feel for what scouting at regional competitions would be like. It was there that I found a problem with transferring via bluetooth.

Previously with our USB fobs, we renamed each fob to match the name of the tablet that was being scouted, so renaming the file that came from that tablet was pretty easy. With bluetooth, we didn’t know which file came from which tablet, so we had to manually open each file and find the device name in the data in order to rename the file.

I wasn’t really a fan of this process, and luckily I had been learning Python through the school’s entry level computer science class. So, I modified the original .json to .csv conversion script (RStoCSV) to find the name of the device and rename both files.

From there I decided that manually converting each file was also too hard, so I made it so you can select a folder and convert + rename all the .json files in that folder. It was a simple project, but it was nice to take what I had been learning in one of my classes and actually use it for something productive. You can find my modified version here!

Also, at Cow Town, I got to be our team representative for alliance selection for the first time. I was kind of nervous at first, but it ended up being a fun experience. Looking forward to doing it this season!

Post-Kickoff!

Nothing too exciting is happening in the strategy department yet. I’ve spent the last 2 days reading (and rereading, and rereading, and…) the game manual to see if I can find anything interesting. I’ve been compiling resources such as field maps and cheat sheets that we might want to use, which are all in the All the CHARGED UP resources post.

I’ve also been looking at mechanisms from the following past FRC games:

• 2011 Logomotion - picking up inflatable game pieces
• 2012 Rebound Rumble - balancing on teeter-totter type of thing
• 2019 Deep Space - handling picking up two game pieces

as well as looking at this season’s FTC game, POWERPLAY. I volunteered at a couple of FTC events during the FRC offseason, so I’ve seen some cone-manipulating mechanisms in person (albeit on a smaller scale).

Strategy/scouting wise, we haven’t talked about much of anything yet. My in-person time has been spent working with the Mechanical team for any ideas (and checking the rules a bajillion times to make sure they’re legal), as well as working with the Programming team for our autonomous routine paths.

I think that’s all for now, hopefully by the time I make another post we’ve unearthed something interesting! Good luck to everyone this season!

[DChilson]

We will have additional motion on the manipulator that will allow us to avoid interference. This is why we are leaning away from the single stage elevator, there isn’t sufficient space in the starting configuration for any of the manipulators we vaguely described in our meeting. Once we prototype these this week, we will have a better idea of what volume we need and possible ranges of motion so the vertical pivot point can be refined.

[Bmongar]

1/14/2023 - First build Saturday went well.

Programming
We made some updates to our SDS Swerve library. We have it working for the Mk4 modules, We had tried a couple of different Chief Delphi solutions which seemed to leave our wheels not steering. We went back to our old code and added position support to the module and it seems to know where it is correctly! Our code was old enough it didn’t have MK4i support so we have added a yet untested Mk4i helper class in preparation for things to come. The swervelib code can be found in a different repo as we were using an imported version of last years bot. To avoid chasing compile errors we removed all the neo code so we didn’t have to fix those method signatures.

We scavenged parts from an old practice bot and put together a test board so we can test vision coprocessors in their native environment.

We discovered that an adjustable power converter that @SLAB-Mr.Thomas had recommended to me continued to output 12.1v even when the power had dropped to 6v. It also retained its settings and comes on on power supply.

We also played with the LL2 with the 2023.0.1 firmware with an attached Coral Edge TPU we acquired last season for our Axon experimentation. We had believed based on another thread that the preloaded model would identify cones and cubes however the default model knows many objects but the game pieces are not within them. We also currently see no way to add or train a new model.

We also attempted to calibrate Photon vison but found it hard to calibrate without almost filling the entire screen with the checkerboard image. We are not sure if this is a bug and have a question open on the Photon Vision release thread. We stopped experimenting with it at beta5 and it looked good and was easier to calibrate.

Programming still has a long ways to go.

Mechanical
Mechanical had a busy day.
They converted our spare Mk4 to a Mk4i. They would have done the rest of them but those darn programmers were making sure swerve would work.

We did some intake testing based on 1339’s roller intake and decided we want to pursue that farther. We do need to determine a mechanism of positioning the angle of the rollers that doesn’t add too much weight or complexity.

IMG 4124

Mechanical did test the tolerances of our new Voron 2.4 350mm and found it off only .0005 inches on a 3 inch side.

We are still planning on using the Thrifty Bot telescoping arm in a 2 stage configuration despite concerns about block friction. To address that issue we designed some rolling bearing blocks for both the 1 inch tube and the 1.5 inch tube

The 1.5 inch has 4 bearings where as the 1 inch only has 2

Our test print will be ready in the morning.

The carpentry crew was busy building the grids.

Also the current mechanical design is a small bot, focus on maneuverability. We are looking at 22 inches square. That will present some challenges packaging everything in such a tight space. It will allow more room on the charging station and some larger tolerances shooting the gaps on the side of the charging station while cycling.

Monday we officially plan to retire Sidewinder and use its components to complete the robotic circle of life. There may be an observed moment of silence sometime to commemorate the bot.

If you have any questions, concerns, or advice let us know. Answers from @Embily , @alynn , @tylurr , @deary , @Nora_C , @saatb , @MrKrabs , @DChilson can be considered official.

[Woolly]

How are you planning to manufacture the bearing block? We actually managed to break the 3D printed blocks that came with our thrifty bot telescoping tubes by dry-firing them (instructions? what are those?), so maybe a hubs.com print is justified for that part.

[Bmongar]

We had planned on testing with our printer until we are satisfied with the quality then sending off to get printed in onyx. Last year’s climbing was especially violent and hard on blocks. We feel with the light load, controlled spooling, and usage mostly in protected zones onyx will suffice. But testing will occur.

[jedgar6]

maybe steal from:

Making New End Blocks for the Thrifty Telescoping Tube

You can find the cat model of this part in our 2022 robot in our OA blog from last year. [PXL_20230114_141915575.MP] [PXL_20230114_141901936.MP]

[Nora_C]

Hi I’m the public relations/ spirit/ awards lead on team 5013. I am going to divide my updates into what the three of those subteams have been doing.
Awards: We plan on submitting for all three FRC submittable awards (Dean’s list, Woodie Flowers, and the Impact award) In week one the awards team has finished the executive summary outlines. Last Tuesday, all of 5013 discussed, voted, and decided who to nominate for the Woodie Flowers award. Next week the awards team hopes to complete a thorough outline for WF and start an outline for the impact award essay.
PR: This week we made plans for more student involvement online. We plan to have regular updates on our website: https://www.trobots.org/ and on our social media: Instagram: @ 5013trobots Facebook: @ PHHSTrobots
Spirit: Development began on our seasonal logo before kickoff. Our logo for Charged Up should be done soonish. We hope to have it done somewhat early so that our apparel has plenty of time to arrive.
I think that is all. If you have any questions or want further clarification, feel free to ask. This is going to be a great season and I can’t wait to keep you guys posted!

[DChilson]

We have finished another weekend including a bonus build day and made fantastic progress. Our primary goals for this weekend were to finalize our frame dimensions, convert the Mk4 SDS modules to Mk4i, and get the CAD to a place where we can finalize last minute parts orders.

1. We have had many conversations with the programming and strategy sub-teams and finalized on a 22" square robot perimeter. The students believe there is more upside to a small, agile robot than additional space to package our components. The design team has already updated the CAD to reflect this decision and worked on repackaging the electrical systems.

2. The mechanical team completed the tear-down of the Mk4 SDS modules and converting them to Mk4i. The team went ahead and completed inspections of the components as they went and determined the primary swerve bearings and tread needed to be replaced. We now have a full set of Mk4i modules that are ready for another season of play.

3. The design team is proud of our current CAD progress. Our team has always fallen short on the CAD side and we are working hard to get a quality product out on schedule for the mechanical team. We don’t have the driving mechanism for the arm extension or the manipulator modeled yet, but we are going to finalize the last minute parts orders. This is the furthest our designs have ever been at this stage in the season and we credit the push on CAD as well as our Open Alliance commitment for this.

We finished the 3d print on the bearing blocks this morning and realized we made a rookie mistake and mixed up the outside dimensions with the inside dimensions of the square tube arms. We fixed our error and printed new blocks with time to spare. Our new 3d printer is performing beyond our expectations and holding tolerances we’ve never experienced before. The bearings fit perfectly and we look forward to getting these blocks installed.

Wire management for the arm is still undecided. We have discussed several options but have not finalized on a decision. We’ve reviewed a few coiled cable options but are still searching for the right solution, everything we’ve seen so far isn’t quite what we need.

We have an ambitious week planned and look forward to making some serious progress.

1. Complete CAD of robot minus intake
2. Build drivetrain
3. Mount plywood a-frame
4. Wire electrical
5. Build base stage of arm

Onshape

Our mechanical team converting the SDS modules to Mk4i under alumni supervision.

Our mechanical team converting the SDS modules to Mk4i.

[saatb]

@DChilson

We have had many conversations with the programming and strategy sub-teams and finalized on a 22" square robot perimeter. The students believe there is more upside to a small, agile robot than additional space to package our components.

To expand on this:

Our initial discussions about frame perimeter were already towards the smaller side. We didn’t want to be bigger than the Everybot (22" x 28.3"), and we believe the advantages outweigh the disadvantages enough to justify this decision.

Some of these advantages are:

• being able to squeeze past defense
• fitting 3 robots on the Charge Station
• easier to fit in the small gaps to get into our Community

One of our comparisons to this was designing your robot to make it through the Trench Run in Infinite Recharge (something our team did). If you wanted to fit under there, you couldn’t do all the fancy things you could with a taller robot. We think that the same idea applies here.

In the end, we decided on a 22" square, which gave us enough space to package our arm mechanism and all of our electronics. We’ve accepted that this means we won’t be able to add any new fancy mechanisms, which is a risk we’re willing to take. This is also a way for us to explore our limits and see how we can make this challenge work, since this is by far the smallest robot the team has designed.

[Bmongar]

Brief programming update since @DChilson handled our mechanical heavy update after Monday’s build.

We copied last years drive train and utilities over to the new project. We reorganized the project somewhat. We put Constants into their own folder. We are going with a constants class per subsystem to minimize the merge conflicts. However CANConstants is still shared as it makes it easier to know what CAN IDs are used.

We starting the work to convert to using PathPlanner’s’ SwerveAutoBuilder in our factory

We have also made some place holder classes to get things started.

The upcoming work today will be to focus on calibrating the cameras we plan to use for PhotonVision.

[EvanM]

Is the cad arm the demensions you are planing to go with? One problem i see with those dimensions is that you wont be able to operate the arm outside of the community and loading zone casue the arm will be outside of the frame perimiter in both the “front” and “back”. Is your strategy to go straight to the loading zone and ignore the per placed game peices?