FRC 6328 Mechanical Advantage 2023 Build Thread

Kickoff Plans 2023

For 2023, we decided to take a step back from our approaches the last couple years and try to develop a more streamlined, prototype-oriented, approach to the first weekend. Many ideas present here are borrowed from @ShelbyA’s kickoff resource. I recommend that anyone and everyone looking for how to run their kickoff tomorrow review that presentation. It is unmatched.

We are making use of the 2023 version of our kickoff worksheet, which helps the team to fully understand the rules in the first few hours. We have streamlined this worksheet since last year as well as updated the examples to include the 2022 game so it is more approachable for current rookies and second year students. This worksheet drew heavy inspiration from the 2791 kickoff worksheet and 1678 offseason strategy training.

Below is our schedule for the first two days. You can expect blog posts from us at the conclusion of both days. I will also briefly explain the important aspects of this schedule below.

pasted image 0

Break into groups, Manual reading, and Worksheet completion

Similar to last year, we are breaking into predetermined groups for the first read through of the manual. There will be separate student groups and mentor groups. Student groups will have a “leader” who is just there to ensure the group stays on track. The groups will read the “relevant” part of the manual aloud, switching the reader every paragraph. We think this will be good for getting a good initial understanding of the manual, as well as accommodating team members who may have a harder time reading. Following the reading, teams will complete the aforementioned worksheet in the order the questions are written. This should hopefully mean that by 3:45p, present members have a strong understanding of the manual.

Robot Archetypes / Game Sims

The easiest way for me to explain is to just post a screenshot of Shelby’s presentation:

We will be mostly following this methodology, but plan to probably spend longer on this exercise. Some students and mentors will start work on potential game simulation tools (details TBD)

Discuss game info, Discuss high level strategy, Robot priority list

At the start of Sunday, we are going to have a relatively off the cuff discussion of what we think we may have missed the first day. It’s hard to know what to expect here. Following this we are going to have a discussion about high level strategy for the robot and what we want to see from the top-most level. This can be things like “be able to solo a rocket” or “be a strong low goal scorer.” Immediately after this we create our guiding priority list, which will be using the MoSCoW format (another idea we stole from @ShelbyA , noticing a trend?). This means that our goals are broken into Must Have, Should Have, Could Have, and Won’t Have. During this part we intend to get more specific and have things like “intake two balls at the same time” or “5 ball auto.” This list will be living and changing over the coming weeks, but will be the guide for our prototyping and design strategy.

Brainstorm mechanisms, Select early prototypes, Prototyping groups coordinate

For the rest of Sunday, we will be coming up with prototype ideas and starting the initial work to get them rolling. Early on, we will break into our preassigned prototyping groups, consisting of 6ish students and a couple mentors. These groups are determined based on a balance of team member experience, subsystem interests, and subteam. These groups will come up with ideas for any prototype for any subsystem and prepare to pitch them to the greater team.The full team will then discuss these prototype ideas and select which we are going to move forward with. The prototyping groups will then reconvene and coordinate for how to move forward.

If you have any questions or comments, feel free to shoot me a PM or post here.

-Connor (6)


Kickoff Day 1

Today for us was pretty much entirely dedicated to understanding the manual and getting a grasp of the strategies we expect to be present. I’m gonna keep it brief because I expect we will have a pretty substantial post coming out tomorrow.

Initial Observations:

The Golden Corral

I’ll start with the main one, that I’ve been surprised hasn’t been getting much coverage on CD until the past lil bit.

From our understanding, G403 permits teams to carry an unlimited number of game pieces from their loading zone to their community as long as they get between them in less than 3 seconds, as that is the defined duration of “momentary.” We don’t feel the blue box is particularly relevant to this, as “egregious” is only assessed following a violation of the rule, which never occurs in this hypothetical.

Above is an example called the “Golden Corral” designed quickly by @Matthew3 which is a robot that could theoretically collect a very large payload of game pieces from the single substation and “drag” them back to its community, all while only having to move at 80ish in / sec, to complete the journey in 3 seconds. This design specifically has its assortment of practicality issues, most notably with retaining game pieces and not getting caught up, but the concept remains.

We’ll see if the rules regarding this stick through the first couple team updates (I expect they will not). It is a very fun concept nonetheless.

Smaller Notes

  • The charge station does not have practical space for three full size robots.

  • Is suction a form of attaching to the field arena? Rule isn’t 100% clear (only 99%), eagerly awaiting Q&A.

  • Scoring visibility is poor due to how close you are to the wall and having to basically look through other driver stations for far grids.

  • Loading station visibility is poor because of small game pieces and big robots and far away.

  • Scoring in your opponent’s co-cop grid could be useful in very niche situations.

  • Intaking cones from the ground could be used to “steal” from our opponents loading zone, or for cones in the midfield.

  • Field seems quite swerve friendly.

  • Field seems to encourage smaller robots.

  • Docking the charge station in auto means we don’t need to worry about fitting three robots in endgame for the RP.

  • The robot can be used to communicate to the human player as to what gamepiece should be deployed

  • This game seems hard.

  • These cubes are gonna pop.

Bigger and better post out tomorrow talking more about our strategy and prototyping




Mini Software Update #1: Field Constants & Coordinate Systems

Wow, what a thrilling way to kick things off. /s

Before we get too deep into software development, we want to take a moment to share our updated FieldConstants class for 2023. It has a collection of useful constants describing the layout of most elements on the field, including:

  • Dimensions and shapes of all regions (community, charging station, loading zone, cable bump)
  • Positions and heights of all nodes for alignment
  • Positions of all staged game pieces for auto
  • 3D AprilTag poses

Last year, we used our 2022 field constants class to define trajectory waypoints in exact poses relative to specific field elements, and we plan to do the same this year.

The 2023 field is particularly annoying interesting because it’s mirrored rather than rotationally symmetric. We’ve opted to set up our coordinate system such that the origin is always at the right side of the blue alliance wall, regardless of our own alliance color. The FieldConstants class defines all of its translations for the blue alliance (e.g. blue community and blue loading zone). It includes a method (allianceFlip) that will check the current alliance color and flip Translation2d and Pose2d objects if necessary.

Tentatively, our plan is to generate all of our trajectories as the blue alliance (near the origin), then apply the allianceFlip method to the setpoint pose while following. The same basic idea would apply anywhere we need to use a reference pose that changes based on the alliance color (like for auto alignment in teleop). We’ll have more details on those features as we develop them further.

We hope this class may prove useful to teams as we all attempt to wrap our heads around a new field layout.




I imagine this will be a similar situation as the code orange post game traversal. It’ll be a cool concept but still get shut down by FIRST


Kickoff Day 2

On Sunday, we developed our high level strategy, robot priority list, and started working on ideas for prototypes. Like we did last year, we have developed our priority list using the MoSCoW prioritization method. This is a living document that will definitely change, especially in the coming week, as we realize things are harder/easier or more or less valuable than we initially thought. We will come back to this list to help with guiding all of our prototype ideas and design considerations.

MoSCoW Priority List

All items in each group are unsorted. We find that very fine sorts like you find in a lot of priority lists are too granular and you don’t really gain from it in almost all cases. Asterisked items are subject to a greater deal of change than the others, generally because we don’t know how hard prototyping will show them to be. You can click on any item for a brief explanation as to how that item ended up in that category.

Must Haves


Seems important idk

Be Fast

Long cycles and many of them. Potential for defense means we need to slip by

Drive in any direction

Great for alignment, dealing with defense, and autos. Not really any downsides to swerve here beyond maybe getting up the ramp which may require higher bumpers or some other funky stuff.

Have clearance to go over bump and charging station

Important for navigation and auto/endgame points and RP.

Play defense

Good to always have the option. This means being able to retract everything as to not go into opponents frame or destroy ourselves, as well as being robust.

Pick up cones

Not specific here, as we don’t know if going from the double substation or the ground is better yet (accounting for difficulty). This will hopefully be figured out via prototyping in the coming week. We just know we need to do one of them.

Intake cubes from the ground

Seems very easy to intake the cubes, potentially easier to do from the floor than the substation. We need to intake at least one gamepiece from the floor for autos.

Score cubes in the hybrid node

Bare minimum scoring. Can get 3 nodes this way.

Score cones in the hybrid node

Would like the option to score cones there instead of cubes. Shouldn’t take away from other mechanisms as worst case you drop it from a higher height than right above the floor.

Score cubes on the mid node

Opens up higher value, and more scoring locations than just the hybrid node. Also alongside scoring cones mid allows us to score up to 6 links and achieve the RP.

Score cones on the mid node

Opens up higher value, and more scoring locations than just the hybrid node. Also alongside scoring cubes mid allows us to score up to 6 links and achieve the RP.

Indicate to human player what the robot wants (actively)

The human players are far away and we need to tell them what gamepiece we want using tools either on the robot, in the driver station, or in the stands. Purple and yellow LEDs on the robot seems like the path forward.

Automatically engage the charge station in Endgame

Important for RP and also a good number of points

Balance on the charge station in auto

Doesn’t matter for RP but gives good points and with the right bumpers/other solutions is a pretty easy software problem.

Automatic placement of game pieces

Poor driver visibility and fast cycles means we need/want fast, automatic scoring.

Semi-automatic collection of game pieces from double substation

Poor driver visibility and fast cycles means we need/want fast, automatic intaking. Does rely on needing to get game pieces from the shelf however.

Semi-automatic alignment to score

Poor driver visibility and fast cycles means we need/want fast, automatic alignment. Depth perception from the driver stations is really rough.

1 Game Piece + Balance auto

Very achievable and sets us up well for the Activation RP.

Should Haves

Pick up cones from the double substation shelf*

This is one of the two main approaches we see for picking up cones. This one allows for you to know exactly where you are getting your cone from and the exact position it is in. However it means you always need to do full field cycles if this is your only way of getting cones.

Pick up cones from floor in any orientation*

This is one of the two main approaches we see for picking up cones. This one allows for you to get cones from wherever you want on the field as well as pick up your own, (or your opponents) dropped cones.

Score cubes on the high node

Highest scoring potential location. Allows for more links, however at most levels being restricted to 6 links is fine. This is desirable both just for the raw points advantage, as well as being a good teammate for a team that maybe can only do mid/low very well or prefers them.

Score cones on the high node

Highest scoring potential location. Allows for more links, however at most levels being restricted to 6 links is fine. This is desirable both just for the raw points advantage, as well as being a good teammate for a team that maybe can only do mid/low very well or prefers them.

Semi-automatic collection of game pieces from floor*

This is pretty much tracking and automatically grabbing gamepieces. This is very nice especially for when you are intaking at the loading station and it is hard to see for the drivers.

2 Game Piece + Balance auto

Seems relatively achievable and simply scores more points and progresses towards the sustainability bonus more.

3 Game Piece auto*

A higher scoring auto for when our partner is going to be doing the charging station auto.

Could Haves

Pick up cones from the ground (upright)*

Depends a lot on what the other mechanisms look like but this opens up being able to do cone autos, as well as grab cones located in the midfield during teleop. This also allows us to pickup cones that happen to fall vertically from the substation.

3 Game Piece (Link) + Dock/Engage auto*

Going to be tight on time but achieves a link in auto as well as contributes to the climb RP. The engaging part is potentially removable as the difference between docking and engaging in auto does not matter for the RP.

* Can access (intake) game pieces within the hybrid nodes

Allows us to pick up game pieces we may have dropped and had fall into the hybrid node. Also allows us to move gamepieces that were score in the hybrid up to a more valuable node or to complete a link elsewhere.

Score on both sides of robot

This has the potential to save on cycle time, especially in auto, without as many downsides of only being able to score on the opposite side.

Intake game pieces within frame.*

Allows us to “steal” fallen gamepieces from the opposing loading station or community.

Active manipulation of the charging station bridge to force it into a position*

Unclear on usefulness or practicality currently, more testing needed.

Won’t Haves

Pick up cubes from the double substation

Picking cubes up off the ground seems too easy to be wanting to get them from the substation. Potentially adds too much weight to the end of whatever device is holding the intakes.

Score (only) on opposite side of intake

Scoring only on the opposite of the intake seems like it will lead to annoying scenarios where you are repeatedly turning around in the community either when you are moving gamepieces around or when you are picking up dropped ones.

Extra small footprint to fit on charge station comfortably with alliance partners

We found that with two average sized alliance partners, you are looking at a <=24" horizontal dimension of your robot. We feel this is outside our scope, as nice as having a mini robot would be.

Assist one alliance member in climb

In this case it refers to lifting a partner up the climbing station. We don’t see this as especially useful or necessary for the amount of complexity.

Assist both alliance members in climb

In this case it refers to lifting two partners up the climbing station. We don’t see this as especially useful or necessary for the amount of complexity.

Some Notes

  • There is a bunch of potential nuance in the different heights of scoring and intaking from the ground vs the double substation. Such as a robot that can score one or both gamepices low, mid, and high from the substation, but only low from the ground.
  • The Golden Corral (pictured a couple posts up) remains unranked, as we expect it to be canned in a Team Update, or at least we can hope.
  • Autonomous and programming specific considerations are somewhat difficult to position because developing them often takes almost no resources away from the other aspects of the team. Almost all development happens in simulation.


  • We don’t think you need to score on the high nodes to play at a good level. Especially considering a strong high node team will often not value high nodes that much when it comes to alliance selection, as there are a restricted number of them.
  • When it comes to late first picks and all second picks, we think being strong at one game piece is almost exclusively better than being middle of the pack at two.
  • Engaging the charging station in auto essentially does not contribute to the RP, so in matches that aren’t tight, you can potentially sacrifice that to continue scoring game pieces.
  • Reorienting cones from any position seems difficult and we expect to have prototyping both from us and other teams drive how much we prioritize this. If getting it off the floor is slow enough, it could very well be faster to do the full field cycle instead.
  • We are aiming to at the minimum be a fast low-mid cycler with a ground cube intake and substation cone intake. Hopefully we do more, but this is where we will go back to as our base.
  • We think defense will be somewhere between 2020 (somewhat mid) and 2019/2022 (strong). With the lack of pinch points it can be hard to pin teams, but the full field cycles and high visibility means defenders have a good amount of room to play with.
  • Docking / engaging the charging station seems challenging to do with 3 normal sized robots, and is not necessary for the RP. We want to see if we can avoid even wanting to do this.

Moving into Week 1

We are going to be immediately moving forward with full team prototyping efforts. We’ll have some pictures and videos out as soon as we can. Our initial focuses are primarily cube floor intakes, cone floor intakes (and subsequent handling), and cone substation intakes. We have a few ideas for multitools that can do both as well.

Things we are looking at:

401 in 2019: Elevator on a pivot

233 in 2011: Telescope on a pivot

1619 in 2018: Full width intake on pivot

1684 in 2019: Dual position intake for two game pieces. Similar version could potentially work here

2011 Logomotion Matches: We expect a relatively similar field dynamic to 2011

Other OA threads: Be sure to check out 3467s and 3847s fantastic posts

Ri3D Redux: Most notably this post about swere on the charging station.

If you have any questions or comments, feel free to post here or shoot me a PM.

-Connor (7)


Week 1 Prototyping woooo!

The past few days we have been in full swing with full-team prototyping. We found HYPE blocks and ProtoPipe to be a great solution for quick and easy assembly and iteration of prototypes, although we had some issues with consistency. We played around with a few different ideas, including a 2018-style intake with side rollers, Kettering University’s RI3D intake, and an intake designed for the flange of the cones. Here are the designs we found the most success with:

2018-style with side rollers:

  1. This prototype uses two tiers of compliant wheels to grip game pieces. The outer wheels intake the cube, and the inner wheels intake the cone. We found this works fairly well for both the cone and the cube. A downside to this approach is that we can only pick up cones in the standing orientation.

Cube intake: 6328 Dual Layer Prototype (Cube) - YouTube

Cone intake: 6328 Dual Layer Prototype (Cone) - YouTube

  1. We also made a different version of this which just uses 2 straight lines of rollers that would theoretically actively actuate in and out to intake cones or cubes. This one showed some promise but had reliability issues so there isn’t much incredible footage.

  1. For a more robust iteration of both these designs, we designed a quick CAD model which we are cutting out of 1/4 inch MDF. The idea with this version is that the positions of both sets of rollers (mounted on the small plates) can be adjusted on the main plates of the intake to find the optimal placement and compression of the game pieces. The top and bottom plates have mounting for Versa or UltraPlanetaries, so we can power it using our motor test rig instead of drills. We will have lots of videos/data about this one out as soon as we can.

Double roller cone (and cube?) intake:

This prototype is very similar to 3847 Spectrum’s design which uses two sets of rollers placed closely together. We built this to grab the flange of fallen cones, but after seeing Spectrum’s similar design intaking the top of the cone we tried that too. At high RPM, we found that it’s even able to grab the flange of cones parallel to the intake (3rd video). The current idea we have in mind for picking up cubes and standing cones is a pivot on one of these rollers that can open up and intake/outtake from above the game pieces. This is especially useful because we found cubes scored from above the shelves rather than directly in front of them might be less likely to bounce off. As it currently stands, the rollers are quite heavy rubber from our old 2020 intake, but we may explore alternatives such as thin walled aluminum tubes wrapped in grip tape (similar to some 2022 intakes). We are also going to be trying different types of wheels, more specifically compliant wheels and 3d printed vector wheels. This design seems promising and we will be devoting more resources to it.

We have also made some progress in thinking about what the overall superstructure of the robot may look like. 3467 made these graphics to illustrate potential robot archetypes. We are big fans of both the double jointed arm and virtual four-bar on a slanted elevator and are actively looking into how these may fit into our intake designs and feasibility of making them.

We have some more ideas in progress that we will get out later this week. If you have any questions or comments, feel free to post here or shoot me a PM.


–Advaith, @aryan , @manthan , @Matthew3


Can you elaborate on this?


The RP requires 26 charge station points. In auto it’s 8 for docking and 12 for engaging. If you engage in auto there’s still 14 points left for endgame which means you need 2 engaged (20pts) or 3 docked (18pts). If you dock in auto the threshold increases to 18 but either way the requirement is the same 2 engaged or 3 docked


Understandable but the described strategy is being used for “matches that aren’t tight”.

Since they aren’t worried about the score being close is the idea that the time difference between docking vs docking and engaging is such that they can get the extra auto cycle off thus be closer to the Sustainability Bonus?

**Edited due to similar profile pics making me think i was responding to OP my bad @Connor_H


To be clear, I (OP) am not @TheGamingR0BOT. Both your assessments are correct. While perhaps small, the time savings in not having to worry about balancing the charging station in auto may be the difference between hitting a second or even third gamepiece in that period, pushing you closer to the Sustainability Bonus.



similar looking pfp, but different person.

edit: oops, Connor already clarified.


What size roller do you guys use and how far away are they from each other?

1 Like

Very cool design! Are you thinking of picking up cubes with the “cone pinch” design as well? Or 2 “separate” Intakes?


The roller has a 1.75" diameter and 22" length. Again, this might be a bit longer than what we would ideally put on a competition robot, as these are repurposed rollers from our 2020 robot. The edge of each roller is 1" apart and the centers of the rollers are 2.5" apart.

With the “cone pinch” design we are thinking of intaking cubes with the same set of rollers. Some options we are thinking about is putting one roller on a pivot so it can open up to be wider, as with their current compression the rollers cannot intake cubes. Another idea is letting the cube run under these rollers and go over the bumper into another intake/grabber that would be mounted onto our superstructure. This second intake would be what scores the game pieces.


Prototype Progress!

Over the past few days we have been hard at work iterating on our intake prototypes. We continued developing our side roller and flange grabber prototypes, as well as explored some new ideas.

Side Rollers

We assembled our MDF prototype, which turned out to be quite large and hefty, and used a ratchet to mount it to our swerve drive base for testing. Although we were skeptical, the ratchet strap worked surprisingly well for keeping the intake in place. Initially we used a motor test rig to control all four motors, which inevitably led to quite a sketchy setup.

We eventually switched to running the motors straight off of the robot. With the adjustable design, we spent a good bit of time modifying the position of both sets of wheels and the back stop, as well as trying out different types of compliant wheels and playing around with the intake angle. The intake seemed to work well for the cone, but we couldn’t get the wheels to hold on to the cube after intaking it. To fix this we made the outer wheels constantly spin at a slow speed, which we found not only keeps the cube in place, but also helps guide the cone into the back wheels. We also swapped the outer set of wheels to smaller, more compliant wheels, which worked better with the cube. We were eventually able to run cycles with the robot, and the intake seemed to work fairly consistently.

With this design showing promise, we have started designing a next iteration made out of polycarb, which will be smaller and lighter than our current version. We are planning on putting the polycarb intake on the end of an arm that would attach to the swerve drivebase, and would be able to score both the hybrid and mid nodes.

Flange Grabber

We improved this design quite a bit over the past few days by finding ways to vector the cones into the middle of the intake. First, we added 3d printed mecanum wheels on the front roller and covered the back roller with duct tape because the rubber was too grippy and didn’t allow the cone to vector. We found that this worked pretty well.

For the next iteration we fully replaced the back roller with an ABS 3d printed corkscrew to help vector while gripping the feet of the cone better. We also slightly offset the rollers to the front one is higher than the back. The main problems we found with this roller was that the edges of the corkscrew were too sharp and the pitch was too low. Also, mounting it on the robot cart made it a little higher than we wanted. Overall, it was decent.

Attempting to solve these problems, we improved the corkscrew by increasing the pitch, blunting the edges, and increasing the diameter. We then CADded side plates and cut them out of MDF so we could keep the intake at an ideal height and wouldn’t need to use the robot cart anymore. This design unfortunately did not work very well as the compression of the flange was too high to work reliably. It only ended up working about 20% of the time, and was shredding up the cone a little.

To solve the problems of the previous iteration, we CADded and cut new plates that reduce the compression, slightly lower the rollers, and add space for another roller of 4” compliant wheels to go below the corkscrew. This additional roller is intended to help kick up the cone into the top two rollers. It is not built yet, but we have our plates cut and ready for Monday.

Other Prototypes

We played around with something similar to 4481’s “weed wacker” prototype to try to orient the cone in the same direction no matter its position. As expected, this worked best when approaching the cone straight on, and the effectiveness tapered off as the angle of the cone increased. This could potentially be something that is placed in front of an intake, to orient the cone in a position that the intake can pick up.

We also started working on a claw-style intake with 2 positions, one for a cone and one for a cube, using parts taken from the 2018 intake design. We started only using one set of wheels to prove the concept and had temporarily placed some surgical as a backplate so that the game piece doesn’t fall out. We played around with different piston sizes in order to find one that is the correct length and stronger than the tension of the surgical tubing. We found that this intake did a good job at holding on to the game pieces, although we may need to test out different compressions so that we can intake a game piece earlier. Next time, we want to add a second set of wheels and also mount this to the robot and test out different heights.

As always, if you have any questions or comments feel free to post here, or shoot me a PM.

-Aryan, @Advaith , @manthan , @Matthew3


First, WOW

And second, that one cube in the gigaintake video looks like a mummy with all the tape on it.


How much compression on the base of the cones was necessary to pull them in? Also is there enough grip to flip the cones to an upright position with that?

1 Like

Do you guys think you could test the vectored intake wheels on the bottom roller similar to how Spectrum is intaking tipped over cones?

Thanks! - Ethan


What was the back roller made from? Is it duct-taped pool noodle?


I assume you’re asking about the flange grabber design, but please let us know if you meant something else. In our first three iterations (two rubber rollers, one vectored + one duct tape, and one vectored + one small corkscrew) we had a 1" space between the edge of each roller. This worked pretty well, but we figured we could use some more compression, so for the next one we tried 0.45". Unfortunately this was too much compression and it didn’t work well, so for the next iteration we are trying 0.7" (not yet tested).

Flipping the intake up with the cone wasn’t tested before, but now that you asked we probably will. The 1" compression was pretty tight on the flange, so any compression of 1" or less we think could definitely hold the cone upright if the intake was pivoted 90° up.

This is definitely something we thought about. The main reason we are using the vectored intake wheels on the top roller is because this design is more suited for when the flange of the cone is toward the intake, not the tip. This matters to us because the corkscrew deals with the bumpy feet of the cone well and the vectored wheels work with the flat face of the flange well. If our next iteration of the corkscrew doesn’t work as well as we hope, using two sets of vectored wheels is definitely something we might test.

It is the same rubber roller from this post but just covered in duct tape. It is much heavier and less compliant than a pool noodle. Because of the weight and size of the rubber, we wouldn’t want to use one of these on our competition robot this season, but they are great for testing.