LigerBots FRC 2877 OpenAlliance Build Thread 2022


Welcome to the Ligerbots 2022 open alliance thread! We are excited to join this great community of teams this season. We appreciate the opportunity to learn more from other teams and hope that this thread can be used as a resource for anyone looking for new ideas.

Overview

The LigerBots, team 2877, is composed of members from both Newton South and Newton North High Schools in Massachusetts. We are a completely student-led team and have over 70 members. Our largest total yet!

We have spent the last few months training our new team members for this upcoming season. Even though it is unlikely that we will have a normal season, we will do our best to adapt to the COVID pandemic.

Schedule

As of now, our plan is to meet Mondays, Tuesdays, Thursdays, Fridays, and Saturdays. Right now we are meeting virtually for kickoff and three-day design, but we hope to be back in person soon. We plan on meeting as much as we can in a safe way.

Open Alliance

Our goal is to post daily updates during our build season. We will compile summaries of the day from each build team with pictures of our mechanisms and links to our online resources. We hope that this thread will show the complete picture of our design process.

Links

Whitepapers:

Distance Learning with Zoom: LigerBots Distance Learning Whitepaper

Vision Processing: LigerBots Vision Whitepaper

Camera Latency: LigerBots Camera Latency Whitepaper

Electrical Test Bench: LigerBots Electric Testbench Whitepaper

Training materials:

FIRST Robot Mechanisms: FIRST Robot Mechanisms.pptx

Drive trains/chassis: Drive trains and chassis.pptx

Robot Recipes: Robot Recipes.pptx

Intro to Robots and FIRST Strategy: FIRST Robots and Strategy.pptx

Sensors: Sensors Presentation.pptx

How Motors Work: Motors Presentation.pptx

LigerBots Build Season Process: Ligerbots build season process 2021.pptx

LigerBots branding standards: LigerBots Branding Standards

LigerBots Chairman’s booklet from 2020: LigerBots Chairman’s Booklet 2020

LigerBots Sponsor benefits flyer: LigerBots supporter benefits flyer

Public LigerBots code repos: https://github.com/ligerbots/

11 Likes

Thanks for doing this, LigerBots! I will definitely be checking this thread.

And thanks also for sharing your training materials, which are some of the best out there in my opinion. Good luck this season!

1 Like

Chassis CAD Release

During our preseason, the LigerBots have been working on a West Coast Drive CAD project that the team will use during the upcoming season. In previous years, we built our chassis out of 80-20, but during the pandemic, we built a successful offseason box tube WCD. Moving forward, we will continue to design with box tubing as it has the potential to lower our robots weight, can be customized, allows us to fasten with rivets, and forces us to design precisely.

Our WCD frame is made of ⅛’’ box tubing, is currently geared in a 11.25:1 ratio at 12 ft/s, and uses four Falcon 500 Motors. Here are some more design decisions we made.

  • 6 inch nitrile wheels: In previous years, we have used Colson wheels in our drivetrains. We have heard from discussion on Chief Delphi that nitrile wheels perform better than Colsons for autonomous driving, so we want to try using nitrile treads this year. We designed with 6’’ wheels in mind in case there are small field obstacles on the floor.

  • Chain-in-tube: We use #25 chain inside our side rails, on 17 tooth Andymark double sprockets. Despite a trickier assembly and repair process, tucking away the chains means more space for mechanisms.

  • Low gearboxes: If they are spatially compatible with our mechanisms, we will use “low” gearboxes that minimize weight and height. In the event that the space currently occupied by gearboxes in the CAD is taken by a mechanism, we also designed over-the-wheel gearboxes as a backup.

  • Bumper mounting: For bumper mounting and the frame perimeter, we had 2 main goals: maximize mechanism mounting space and keep the chain-in-tube as accessible as possible. We ultimately decided on placing sliced 1/16’’ box tube along the side rail, which would rivet into the chassis, and support an aluminum angle that runs the frame perimeter length. For attachments, we have latches on the front and back rails. We used them in our original tube WCD project and an offseason swerve drive and found them easy to use.

This is the link to our CAD: Onshape

3 Likes

The LigerBots begin build season with a 3 day design process. Our team splits into 9 groups of 8-10 students (we have a huge team this year). Each group designs a game strategy and robot. Subteams such as CAD, electrical, and programming, are distributed among the groups to diversify the ideas generated and take into account a variety of technical backgrounds.

After 3 day design, the team will regroup as a whole and choose a robot design to pursue through a “six hats’’ process which reviews each design proposal made by the 3-day design groups. In this process, there are six “hats” which represent different ways of thinking about each potential robot and it’s mechanisms. The 6 “hats” we use are:

  • The red hat, where team members give design feedback using feelings and intuitions
  • The white hat, which calls for facts and information that are known about the mechanism in question
  • The yellow hat, which is all about optimism and what could work
  • The black hat, which explores risks and difficulties
  • The green hat, which focuses on creative extensions of the proposed idea
  • The blue hat, which manages the thinking process.

The six hat process is meant to review each design thoroughly and comprehensively and get the team on the same page about what we envision for our robot based on the risks and rewards of the designs.

The purpose of our 3 day design process is to obtain unique and different robot ideas, so we prohibit any sharing of ideas between members of different groups until the 3 day period is over. As such, we will publish every team’s proposal and the six hats document after our meeting tomorrow.

2 Likes

End of Week 1

COVID Update: In order to make sure that all members of our team are socially distanced, each subgroup will only bring 5-6 students to each meeting.

Our blog posts have been delayed this week so we have a lot of material to cover.

On Tuesday, The Ligerbots began our six hats process described above. Each group presented their mechanism and strategy ideas over Zoom, and the team wrote comments from the six different perspectives.

On Wednesday, build leaders met to narrow down all of the mechanisms and strategies proposed during 3 Day Design into one robot idea that the team will pursue.

Strategy and Priorities

Require

~13 ft/s 6 Wheel Tank Drive

  • Cannot be easily pushed while shooting
  • Familiar design (chain-in-tube WCD offseason project)
  • Swerve modules are backordered to February

Reliable Upper Hub Fender Shot

  • Outscores low hub robots with good accuracy
  • Could be designed to also score in low hub if desired

Upgradeable Mid Rung Climber

  • Increases endgame ranking point odds
  • About triple the scoring efficiency of cycling
  • Leaving upgrade path (to high/traverse) open enables substantial improvement for later events

Desire

High Rung Climber

  • Very high endgame ranking point odds
  • Triple the scoring efficiency of cycling

Lower Hub Scoring

  • Better odds for cargo ranking point
  • Balls get recycled faster than in the upper hub

Aspire

Traversal Rung Climber

  • Points
  • Coolness factor

Robot Design

For a shooting mechanism, the team’s 2 main ideas are a hooded centripetal shooter or a catapult. The hooded centripetal shooter has the potential to give us maximum control over our shots in any field position with a turret and/or a hood, and we would like to be able to control the spin for optimal shots. However, we cannot guarantee the consistency of a centripetal shooter, which is why the other main shooter mechanism proposed was the catapult. Historically, our team had high accuracy with our catapult in 2016 and less precise shooting with a centripetal shooter in 2020. Concerns with the catapult were a lower cycle time since balls cannot be shot continuously, as well as ball storage and overall space taken on the robot. Ultimately, the team will prototype both types of shooters, and see which performs best with this year’s game piece.

Next, the team discussed intakes. We debated on through vs over the bumper and decided to go with over the bumper. It reduces manufacturing since the bumper does not need to be cut, and intaking at a lower height with a through the bumper intake is not necessary since we are not worried about being height constrained in this game. An over-the-bumper intake will also be faster since time is not taken to align the ball. We decided to shoot away from our intake because it allows us to easily line up for the fender shot.

Finally, we got to climbing. Everyone agreed climbing to the mid rung was a requirement, due to the scoring efficiency and endgame ranking point potential. We further required that our robot has an upgrade path to a high or traversal climb. Designing our mid rung climber with this in mind will open up options to significantly improve our robot for later events.

Six Hats Design Highlights

Splitting the team into 9 groups during 3 Day Design allows all ~70 members on the team to participate in the design of the robot. Giving feedback on each other’s ideas through the six hats process helps the team reach a consensus. Here are some of the cool ideas that came out of our brainstorming.

High rung climber, roller intake, centripetal shooter with lower and upper hub scoring capability

Double catapult with a telescope + fixed hook (not shown) traversal climber. The hooks would be spring-loaded.

Similar traversal climb except with an elevator perpendicular to the frame. While these two designs only have a single actuator (the elevator), in the six hats process, people expressed concern about the 120-degree swing into the alliance wall after releasing the high rung. Link to slides

Elevator + Pivoting hook traversal climber

This design is very similar to Ri3D Zoukeeper’s climb and 1986 Team Titanium’s 2013 climb and is proven to work.

Whiteboard Pictures

We have also been working on prototypes, CAD, and other areas and will release blog posts about them soon.

1 Like

Climber Update

This will be an overview of the work we have done towards our robot’s climbing mechanism. On Wednesday we settled on strategic priorities and the climber design was decided on Friday. We have been working on CAD since then.

Strategic Priorities

Demand: Mid-rung climb, upgrade path to high

Prefer: High rung climb, upgrade path to traverse

Aspire: Traverse climb

Due to these priorities, we will CAD a traverse capable climber, but build and test the mid-rung components before trying to upgrade to a traversal climb.

Design

Our design is inspired by 1986 Team Titanium’s 2013 climbing mechanism and the Ri3D Zoukeepers climb. The idea is to line up the elevator with each rung by rotating the robot with hooked arms. The elevator has only one stage and is fixed to the frame. Here is how our design differs from others:

  • Two telescopes over one wider elevator
    • Doesn’t get in the way of our ball path through the center of the robot
    • Simpler to design and manufacture
    • Potentially lighter
    • TheThriftyBot kits available
  • Elevator and arms close to the center of the robot
    • The closer the arm pivot is to the center of gravity, the easier it is to rotate the robot
    • Robot swing is minimized with elevator hooks close to the center of gravity

We like this design because it has almost no swing (Zoukeepers climb). This is because as the elevator pulls the robot towards the traverse rung, the arms on the high rung can rotate and bring the robot’s center of gravity closer before releasing from the rung.

CAD

After a good experience with dividing robots into separate documents for each mechanism while participating in the 1678 Citrus Circuits Storm Surge CADathon, we decided to do the same for our 2022 build season.

Why we switched:

  • All high-level relationships are defined in the main document sketches
    • Clearer design intent
    • Easy to make important changes
    • Inexperienced designers can work in isolation
  • Layout sketches are derived into separate documents for each mechanism
    • Forces everyone to create versions that can be reverted
    • Changes in the main document are added when we choose
  • Assemblies are imported back into main document
    • Main assembly reflects the physical robot (thanks to version control)

We start every mechanism with layout sketches. Here is an explanation of each sketch we made for the climber.

Climber Top

This is the first sketch, which determines the placement of the telescopes, arms, and gearbox. The gearbox plates will be bolted to the drive rails.

  • Orange: 1.5x1.5” Box Tube
  • Green: 1x1” Box Tube that slides up and down
  • Blue: Space allocated for ¼” gearbox plates
  • Yellow: Space allocated for gears
  • Gray: Space allocated for motors
  • Brown: Space allocated for the arms
  • Dotted lines above the gearboxes: Mid rung position

Field Geometry Side

This sketch simply lays out the field geometry. The mid rung is aligned above the shadow in the climber top sketch.

  • Brown: Ground and alliance wall
  • Yellow: Low rung clearance
  • Purple: Max robot height extension
  • Red: Mid Rung
  • Blue: High Rung
  • Green: Traversal Rung

Telescope Side

This sketch defines the height of the telescopes. Given a minimum stage overlap of 10”, 2” of clearance for the hooks over the rung, and 2” between the telescopes and drive rails for the winch drums, we can calculate the minimum height of a one-stage telescope. For now we have decided on a minimum overlap of 10” because less overlap allows us to pull the rung closer to our robot, at the cost of rigidity. The next sketch will explain why the distance we can pull the rung matters.

  • Pink: Approximate robot center of gravity
  • Blue: Telescope base stage
  • Red: Telescope first stage
  • Green: Imaginary rung at fully retracted telescope hook position
  • Yellow: Mid rung

Arm Side

This sketch determines the pivot point and length of the arms. The robot’s weight is fully supported by the arms hooked on the high rung as the telescopes hook on the traversal rung. If the arms are too long, the telescope will not reach the traversal rung. But if the arms are too short, they will be unable to reach the rung when the telescope is fully retracted. Further, we want the pivot point of the arm to be close to the center of gravity so rotating the robot is easier.

  • Gray: Robot chassis
  • Yellow: Robot center of gravity (directly below high rung)
  • Pink: High Rung
  • Brown: Traversal Rung
  • Purple: Imaginary rung at fully retracted telescope hook position
  • Blue: Telescope base stage
  • Red: Telescope first stage
  • Dark green: Arm hooked on high rung, supporting robot weight
  • Light green: Imaginary arm in position to hook a rung after the telescope fully retracts

Why can’t we set the arm pivot coincident to the center of gravity? That would make rotating the robot much easier because the lever arm would be very small. Unfortunately, moving the pivot point there means we need longer arms to reach the rung, which puts the traversal rung out of our maximum height extension.

To avoid this problem, we constrained the arm to be equidistant to the high rung and the imaginary rung at the fully retracted telescope hook position. If we set the minimum overlap between telescope stages to 10” again, the arm pivot is constrained to a single line.

Since we want the pivot point to be as close to the center of gravity as possible, we made the line that the pivot point can move along perpendicular to the line connecting the pivot point to the (approximate) center of gravity.

Our plan is to finish the CAD of the telescopes and gearboxes this weekend. We will post more updates on our CAD progress.

7 Likes

We’re off to another successful build season! We have begun to prototype our ideas and hope to finish in the next few days.

Our catapult team began their work by splitting up into two groups: experimentation and theory. The experimentation team worked with our old 2016 robot to learn about the catapult. The shot was optimized to release horizontally, so they are going to brainstorm ideas for a more virtual shot such as pushing the ball further back on the arm. The theory team broke down the mechanics of taking a shot, including parabola angle/slope. They concluded that a more vertical shot will be more desirable for this game. They also researched how a vertical shot would affect the time taken to release the catapult and the expected launch angle. They decided that the furthest shot they will be able to make is from the launchpad, with an average shot on the top of the tarmac.

The electrical team used their pneumatic test bench to power the catapult group’s initial prototype. They also worked on their brushless motor test bench, but there was an error where the radio could not communicate with the Roborio. After this issue is fixed, the brushless motor test bench will be ready for prototyping.

Our shooter team began constructing a minimum viable product centripetal shooter. The frame is built out of 80-20, and will allow the team to adjust the compression of the ball as they test. The goal of the prototype is to learn more about how the ball will interact with the shooting wheels, to find data that will allow them to pinpoint the optimal compression, and to experiment with the relative velocities of the upper and lower wheel. The prototype is coming along nicely, but they are a bit stuck on the correct implementation of the Versa Planetaries. They expect to have a shooting design during tomorrow’s meeting. After a few rounds of testing, they will commit to the design phase and work with the CAD team to narrow in on a more detailed design to prototype more fully.

Our intake team built a mock bumper and started to make a test bench to test different wheel types, vertical compression, and horizontal compression. They cut the sides and drilled the holes, cut a piece of plywood for the bottom, and cut cross bracing. Next meeting they will assemble and start testing.

Our CAD team has been laying out the master geometry for the entire robot. We use Onshape, and have one master robot document where side, top, and front views are sketched of each mechanism which include essential dimensions, game piece views, and field elements. Using the Onshape “derive” feature, we then move the key sketches from the main document into one file for each mechanism. This workflow system has greatly improved the efficiency of our CAD team. In the 2020 season, we made the entire robot in one document which led to long loading times, and lots of people working in the same place interfering with each other’s progress. Now, the amount of data in each document is significantly less, allowing Onshape to run faster. Deriving the same sketches into different documents allows the whole team to CAD around the same geometry and make sure mechanisms integrate properly.

1 Like

We are now deep into the prototyping phase and will soon be able to start assembling our chassis!

Our catapult team continued with the same groups as before with one group testing the 2016 catapult and the other working on calculations for desired angle and velocity when shooting from the launch pad.

The group testing the old 2016 robot ran into some difficulties along the way and worked to solve these in several different ways. First, when we tested the catapult by itself using our test bench, we concluded that changing the position of the ball in the catapult fork changes the trajectory and potentially the velocity. We also noticed that the ball was constantly rolling off of the end. To try to prevent this, we added shaft collars to the end of the rod. However, after testing this, we concluded that this limited the height of the shot— there was a very low, straight trajectory rather than the high arching one we were looking for. We took off the shaft collars and decided to try to extend the arm instead. We used PVC pipes. At first, we had short pieces of PVC, but this wasn’t working well because it had the same purpose as the shaft collar. So we cut more PVC - enough for the ball to sit on. This proved to work well, and after pushing the two ends of the fork together, we were able to make it over the top of the hub. However, when we moved the robot back, it didn’t have enough power to shoot farther.

We later began prototyping their hopper. Our design has 2 polycord loops running on 3D printed pulleys which take the ball from the mecanum wheel intake and transports it to our shooter. To prototype, we cut two pieces of plywood for the two sides and holes for the bearings and attached the shafts and pulleys.

Our centripetal shooter team updated the design of the prototype, changing the assembly of the launcher. There are now a total of 6 wheels used to shoot the ball. We also added two 80-20 bars that can control both the compression and intake of the ball. We then worked on finding important initial data of our shooter. We started experimenting with the velocities of both sets of wheels, finding what speeds cause the best arcs.

Our intake team started to build a rig that allows us to test the pathing of the ball with 0.75in compression and 4" wheels. The balls will travel under the rollers and over a mock bumper we created and are mounting on a dolly.

We later finished the mock-chassis that goes onto a caddy, cut the plates for the sides of the intake, and started to assemble the shafts and the gears and chain.

The climber team is making steady progress in the design stage and has been working through where to position the climber, how we should design our hooks, and other details. We have decided to use a telescoping climber to climb directly to the mid bar, as our team decided to set that as our minimum goal. Once built to completion, we plan on building a pair of arms that would allow us to transfer up to the high bar and then to the traversal bar by rotating the robot, hooking on, and then swinging down to repeat the process.

Meanwhile, our CAD team has been making steady progress in designing the robot. The general concept is to have the intake center the balls with mecanum wheels which will feed into a polycord conveyor belt, bringing the ball to our shooter. Our main CAD roadblock is deciding how balls will be fed into the shooter. We have been prototyping by having the balls come into contact with the wheels at the same time, however, a ramp and wheel do not achieve this, as shown in the screenshot below. We are considering an angled conveyor belt instead to have our CAD match our prototyping so that the data we collect accurately represents our final robot design. Future prototyping will include how we feed the balls into the shooter, as we have found that changes massively affect the shot.

2 Likes

We have been a little behind on getting blogpost material out so this is a bit of a catch-up update.

Shooter Prototyping

We have decided to go for a two-wheeled centripetal shooter that can shoot at different angles for upper and lower hub shots.

Lower Hub Shot: Fender, topspin, 20 degrees from vertical
Upper Hub Close Shot: 3 ft away, backspin, 20 degrees from vertical
Upper Hub Far Shot: Max 25 ft away, 40 degrees from vertical


More testing videos

We have been testing up to this point with 4 7/8" non-compliant wheels.

We finished testing the 4-inch compliant wheels and decided against using them for two main reasons. First, to get the same arch and distance as the 5-inch wheels, we had to use 4/5ths the amount of power. Additionally, when we attempted to feed two balls in close succession, the shots were inconsistent with each other.

Next, we tested the “marshmallow” wheels and found them convincingly inadequate. We worked on installing “stilts” to allow the shooter to merge with our indexing prototype, and this is nearly complete. We had to widen the shooter’s outside profile to allow the mounting, and the front legs will be long enough to allow a range of angles. Next meeting, we will fine-tune the adjustments for the shooter to indexer mount and then begin testing. I believe this will be a fairly involved process, as we have to replicate a higher degree of precision than the prototypes may initially allow.

Finally, due to the possibility of missing Friday and Saturday’s meetings due to snow, we are planning to meet (as is the rest of the team) on Wednesday.

Hopper Prototyping

We built a simple hopper prototype using 3D printed pulleys and polycord belts. In addition, we cut two pieces of wood in an arc for our “flup” design. We were able to test it and it worked well! However, we are now considering prototyping a different design without the flup wheel. If we have conveyor belts on the side feeding directly into the shooter, the balls would not have backspin and might shoot more reliably.

Field Elements


This is our progress on the upper hub so far.

Intake

We tested the intake prototype with surgical tubing attached and we found that 2.5" initial compression worked well enough and gave us good geometry for the rest of the intake. We also found that we didn’t need a third set of rollers so we are going to CAD a final prototyping plate to test the new geometry with 2" wheels and surgical tubing both with and without mecanum wheels because we aren’t sure if we need them since the hopper might be enough to center the balls. We are also going to start prototyping different materials to work as a slapper roller to help intake bouncing balls.

Climber


This is our climber CAD so far. As described in an earlier update, we are using a pair of pivoting arms and single-stage telescoping elevators to climb to the traversal rung. The goal is to first get a mid rung climber and add on the traversal rung capability if we have time. We plan to have a thriftybot telescoping kit for the top slider but to build a custom slider for the bottom of the 1x1 tube. We wanted to have bearings on the front/back faces as the robot will have some rotational forces.

We have spent a lot of time worrying about our climber’s robustness. In previous years, our climbers worked adequately under normal conditions, but unexpected forces at competitions have caused failures. This year, we want to be prepared for when things don’t go as planned.

The scenarios we have anticipated include:

  • Climbing to mid rung with only one hook engaged
  • Getting bumped from behind while climbing
  • Getting bumped from the side while climbing to mid
  • Transitory loads on arm from sudden impacts

The arm is 32" long and the pivot point is 8" away from the expected robot center of gravity (see earlier post for why we chose these values).

Some ideas we have thought of for mitigating these issues are:

“Cush drive” sprockets on the arm

These are used in dirt bikes. The idea is that the sprocket is mechanically linked through polyurethane rubber so that sudden impacts are absorbed. We could mold the polyurethane used in polycord if we wanted to make a custom sprocket.

J plate telescope + arm support

1/16th aluminum tube riveted to drivebase and to the outer telescope using gussets. The gusset on the telescope is a larger one, kind of J shaped, and attached to that (maybe offset with some spacers) is the arc-shaped aluminum plate, which lives spatially between the two climber arms. The rotating arm can use a shoulder bolt to constrain itself into the curve of the arc. It may be tricky to make the arc plate stiff enough to be helpful.


This would strengthen the telescope and arm, which could be helpful if we are bumped while climbing.

Throw the chain under the bus

Driving the arm with a #25 plate sprocket, perhaps with tensioners to make the chain longer, would make it more likely that the chain would break before shafts or planetary teeth.


Because we don’t want to risk breaking shafts, we decided to gear our Neos with a 25:1 maxplanetary followed by a 4.5:1 chain reduction. We would run the Neo at low power for controllability and to reduce the load on the gearbox.

4 Likes

This is a kind of another catch-up update.

CAD Release

Main: We have top-level layout sketches here which are versioned and derived into other documents. This allows all inter-mechanism relationships to be defined in one place so that all high-level changes propagate to every mechanism. Assemblies from each mechanism document are imported back into the main document for the full-robot assembly.

Chuter Linear belts push the ball into a two-wheel linear shooter

Chassis Chain-in-tube 6 traction wheel drop-center West Coast Drive

Climber Fixed one-stage telescope and pivoting arm traversal climb

Intake Passively deployed, non-parallel four bar linkage, over the bumper roller intake

Conveyor Belt Brings balls from intake into chute

Linear vs centripetal hopper

At first, we had designed a centripetal ball path into the shooter (left sketch). Many teams had success with this in 2020, including our own. But after prototyping both designs, we decided the linear hopper (right sketch) was best given our circumstances.

Linear two-wheeled shooter

We were able to get consistent shots by feeding balls such that they touch both wheels at the same time. We think being able to control backspin will be important for minimizing bounce-outs.

Centripetal hopper

This hopper fed balls consistently but with backspin. The polycord belts were driven by 3D printed pulleys.

Linear Shooter + Centripetal Hopper

Unfortunately, we have no pictures of this prototype. We mounted the linear shooter above the centripetal hopper but did not support the ball path with a curved backing. Consequently, the ball was unable to reach the upper roller at shallower launch angles and shot inconsistently at steeper angles. Although this prototype was inconsistent, we cannot rule out the centripetal hopper based on this alone because a real robot would have proper compression of the ball at all times.

Linear Shooter + Linear Hopper

We fed balls straight into the shooter using polycord. The shots were very consistent and we were able to reliably make shots into the upper hub at several distances.

Team Decision

While many tradeoffs were considered before we came to a decision, to put it simply, the centripetal hopper would have needed more prototyping while the linear hopper did not, so in the interest of time we chose the linear hopper.

Although our first iteration probably won’t be able to shoot from different angles, we want to have an upgrade path towards a shooter that can reliably shoot in both the lower hub and upper hub.

In the end we chose to build a fixed angle linear hopper design with a pivot for the hopper and shooter. This allows us to find the best angle through future testing and makes upgrading to an adjustable shooter much easier.

Climber Progress

We finished assembling our telescope prototype and it slides nicely. We wanted bearings on the front faces because they would be under force when the robot is tilted while traversing. We are using 3D printed bushings for the side faces. We will replace the top slider with the ThriftyBot telescope kit once it arrives. We were concerned about access to the bottom slider bolts once everything is assembled, so we plan to add access holes in the outer tube. Additionally, we have each bolt thread into two nuts to prevent loosening.

Drivetrain

We are building a fancy chain-in-tube drivetrain. Historically we have built our drivetrains out of 8020, but improvements in our CNC and CAD abilities are enabling us to construct our robots with more box tubing and rivets.
Like our 2021 offseason chain-in-tube drivetrain, we use 17 tooth Andymark double sprockets and fixed sprocket center-center distance. Unlike our offseason drivetrain, we opted not to include a 0.018" adder because we didn’t want to risk making the chain tension too tight. If the chain tension is a little loose, the chain might scrape the tube but it can’t come off the sprockets.

4 Likes

Intake

We manually cut these polycarbonate intake plates using a printed template, bandsaw, and drill press since our CNC was in high demand. The non-parallel four bar linkage is designed to have a built-in hard stop at the desired deployment position. We will post videos of it running soon.

Climber

We fixed the excessive friction of our telescope prototype by replacing the #8 plate mounting bolts with #4 ones. In the left gif, the #8 bolts scrape against the tilted edges of the constant force springs, making an awful sound and significant drag. The right gif shows the much smoother telescope after we switched to smaller bolts. Presumably, the thrifty telescoping kits that we expect to receive in the next week or two will not have this issue.


Our software team tested our 2020 robot’s ability to adjust its angle using an arm similar to the one we are planning to build for a potential traversal climb this year. One thing we learned is that even if the neo is braking, it is not hard to rotate the arms. This might cause problems if our motors can’t stop our arms from flopping around with every acceleration. We will elaborate on our challenges with the climber design more in the near future.

3 Likes

Today our shooter team finished rounding and smoothing down both bearing plates, centering the shooting wheels, and attaching the bearings to the second shaft. They began to build a sprocket system for both of the motors.
Our hopper/chute team finished cutting the L brackets to two inches, four pieces of 80-20 to attach the back support, a ball guiding piece for the body of the shoot, and the box tube moving directly across the center of the chassis. They also threaded the four hex shafts for the pulleys.

Our climber team attached the arms they cut to the old 2020 game robot to test their prototype. They also finished one telescoping arm, began working on the second. They hope to start testing soon.

Our intake team decided to try and use the flanges on the bearings to hold the intake plates in and support them in the opposite direction. They made the cross beams out of thunder hex because it would be easy to manufacture and put a bearing on for the idler pulley. They disassembled the two sides of the intake and drilled the holes for standoffs between the main plate and the other plate on one side and for the cross beams between the two sides. They also started to reassemble the shafts with PVC spacers. They tested the belts and found that the 104 tooth belt was too long. They plan on trying the 100 tooth belt soon.

The Pursuit of a Swingless Traversal Climb


We used some cool Onshape mates and configurations to visualize our traversal climbing sequence. Who would have thought the pin and slot mate was just what we needed?

Our climber design is inspired by the Ri3D Zoukeepers and 1986 Titanium’s 2013 robot because we wanted to minimize swinging. Our new CAD assembly illustrated that we need to redesign the pivoting arm hook to make it easier to release from the previous rung. We are considering two approaches:

  1. Shallower hook angle
    While this is the most convenient to the design, we risk getting knocked off the rung if another robot bumps into us by accident.
  2. Longer arms
    If we set the geometry right, we can bring the robot’s center of gravity directly below the high rung with the telescopes. Then if we pull the robot even closer to the high rung, the arms will naturally come off the mid rung even if we have a steep hook angle. We haven’t figured out how to make the geometry work for this yet, but the Zoukeepers Ri3D climb demonstrates this effect. We really like how little their robot swings.
4 Likes

Would you be able to provide a bit more info on the intake? We are designing a similar one but want to know how the polycarbonate acts on a full-size version. We haven’t been able to test ours yet due to shipping delays. What thickness are you using? When the robot moves or takes in a ball, is the intake not stiff enough? Too stiff? Just right? Do you have a support beam connecting the two side plates, or are they only connected with the thunderhex that the wheels rotate on? If you find time for a video, that would be great. Sorry for all the questions, I am trying to gather a lot of information because we have gotten really screwed over in our prototyping this year due to lack of materials and shipping issues. I appreciate the help!!

Sure!
In regards to the polycarbonate, it works quite well, remaining in place, and bending instead of breaking. We’re using a 4-bar linkage to deploy it, mounted to a main side plate (all of those being .25" thick). To protect the pulleys and mount the motor, we have a 1/8" plate on one side of the intake spaced 1" away from the main place (face to face distance).

We’re not completely done with our testing, and aim to make a few more improvements before the final product, but the stiffness of our intake when moving to take in balls, getting hit from the side, and colliding with walls we are quite happy with.

Right now for our prototype, we’re using two cross-beams made of 1/2" thunderhex, rigidly mounted to the plate, but in the final version the frontmost cross beam was designed to be swapped out with 1/2" circular tubing instead ( We’re not sure if this is necessary, so we may just use the thunderhex)

Here’s a video of some of our testing with it.
https://drive.google.com/file/d/1NG6fJ-8rtCt5V9hAEstL_QRhfCNQmaKd/view?usp=sharing

3 Likes

Climber Redesign

The biggest problems with our old climber design were:

  1. Accessibility: The gearbox was riveted to both sides of the drive rails, making it very difficult to remove. Additionally, the winch was very hard to access. Assembling and maintaining this design would have been a nightmare.

  2. Hook release: In the position shown below, it is not possible to remove the pivoting hook unless we have a very shallow hook.

image

We redesigned the climber to address these problems and more.

Hook Problem

We didn’t want a shallow hook because of the risk that another robot bumps into us while climbing. In order to remove a deeper hook from the rung, we must pull the robot’s center of gravity past the high rung horizontally. We modified our layout sketches to achieve this, but at the cost of telescope overlap (this relationship is described in more detail in an earlier update).


The below gif illustrates how the new geometry naturally releases the hook. This is effectively swing-free because the robot’s center of gravity will be directly below the high rung as the hook comes off.

As an unexpected bonus, we were able to bring the pivot point almost directly to the center of gravity, making it much easier to rotate the robot with our arms.

Unfortunately, making the arm longer in this way required our telescope to have less overlap. We mitigated this somewhat by moving the telescopes to the outside of the drive rails and lowering them by two inches.

This new design has a more easily accessible winch and can be assembled independently and riveted to the outside of the drive rail. Mounting the Rev absolute through-bore encoder directly to the Maxplanetary shaft is okay because our arm only needs 40 degrees of rotation.

Further, moving the telescopes out and the arms in allowed us to connect the pivoting hooks. This adds rigidity and helps to ensure that the arms will rotate at the same speed.
image

We had a design review of this gearbox and came up with these areas for improvement:

  • Better solution for tying rope to inner tube

    • Don’t use constant force spring bolts
    • Wire / coathanger?
    • Additional holes in slider plate
  • Stronger mounting to drivetrain

    • bring tensioner plate down and attach to both sides of drive rail
    • Be mindful of bellypan (already riveted to chassis)
    • Rivnuts?
    • Riveted 1/4" plate on inside of drive rail
    • Spacers sandwiching the tube
  • Thicker lightening bridges, especially the ones close to the arm pivot

  • Rivet telescope only to the outer plate so it’s more easily removable

  • Interference with drive gearbox (move tensioner somewhere else)

  • Cut bumper support angle into two separate pieces and bolt them to the outer plate

  • Solution for tying the rope to winch

    • We can drill into the 72 tooth gear for space to tie cable if necessary

    • One versahub with 3D printed or metal plate

    • 1/2" Hex shaft as winch

      • More reduction
      • Drill hole through shaft for knot
      • 15-20 wraps of cable
  • Lateral support for elevator

  • box tube going across but below the chute

We are improving the CAD and hope to start CNCing plates soon. In the meantime, our climber prototyping is moving along and we plan to have a testing setup on our offseason drivetrain from last year so that we can give our programmers something to program as soon as possible.


Today our shooter team corrected the issues with the shaft by adding shaft collars and straightening the shaft. They started the angle adjustment mechanism, which is made of mostly 80-20. They also began discussing meta-concepts for the integration of the lead screw angle. Their primary design is the “H-Slide”, which can be controlled dynamically with a lead screw or can be bolted in place for an adjustable fixed-angle shooter.

Our intake team reassembled the intake with Mechanum wheels. They are CADing the plates to mount the intake to the actual chassis. They also added a second set of bushings on the bars mounted to the drive rails because the bushings weren’t long enough to go all the way through. They hope to soon finalize the main plates and mounting for the Mechanum plates and cut hex shafts to the right length.

Our hopper team added additional pulleys for the direct connection between the motor and the figure 8 belt. While testing, they ran into the following problems: the polycord was slipping on the pulleys and was shredding the polycord, and there was so much friction between the pulleys and polycord that the motor could barely spin if at all. To fix these issues, they realigned several shafts. They also printed larger pulleys, cut longer shafts, and added those pulleys towards the front of the shafts on the bottom. They then cut more polycord and welded it into a belt.

Shooter

Our shooter team cut box tubes for the shooter frame. We also cut 4 triangle plates to hold the shooter structure together. As of now, we’re working towards a fixed-position shooter, and focusing our resources on implementing the CAD put together for V2 (different dimensions, lighter). The chute polycord system works well and we are happy with the consistency of our shots.

shooter_test_AdobeCreativeCloudExpress

Intake

Our intake team began testing and found that the deployment worked better with surgical tubing. We added guiding plates alongside the hopper to prevent the ball from going into the dead zone, where the balls won’t move into the hopper. The mecanums center the balls nicely but get jammed if we try to pick up two balls at exactly the same time. We don’t think this will happen often, but our driver will have to be mindful of this and approach clumped balls at an angle.

intake_vectoring_AdobeCreativeCloudExpress intake_jam_AdobeCreativeCloudExpress

Climber

It’s starting to come together! We assembled our prototype climber with manually cut plates to our offseason drivetrain. Our programming team is working on this while our competition robot is being used for cargo-related testing.

20220224_205453_AdobeCreativeCloudExpress

Our prototype telescope has a sticking problem demonstrated above. We think this is caused by the constant force spring catching on the bolts that fasten the top slider assembly to the outer tube.


This is our prototype gearbox. We printed out CAD drawings and cut the plates manually while our CNC was making shooter and intake parts.

Our CAD team is almost done finalizing the climber CAD. Both “final” climber subassemblies are almost fully manufactured and assembled. We will apply black and orange vinyl soon.

We put tube plugs in our pivoting hooks to support the bushing and tube. They have such a snug fit that you couldn’t tell that the tube used to be hollow.

CAD improvements

We put all the bolts and nuts into CAD to better plan how the gearbox will be assembled and maintained. We decided that the purple-colored spacers would be threaded for 10-32, blue spacers would be ¼-20 clearance, and grey spacers would be 10-32 clearance.

Some other subtle but valuable changes include:

  • Tapped a 10-32 hole for ¼-20 because it would have been a nightmare to get the nut in place
  • Threaded spacers (colored purple) made removing a single plate simpler
  • Riveted the telescope to the outer plate so it can be removed by taking out a few bolts
  • Plates were strengthened by condensing lightening holes
  • Split the bumper angle into two parts that mount to either side of the gearbox
  • Expanded a hole so that we could replace the winch motor without removing the entire inner plate
  • Changed the winch to ½” thunderhex with shaft collars for higher reduction
  • Changed the arm pivot to a bushing instead of a bearing to take loads better
  • Added a ½” spacer to the pulley at the bottom of the telescope because if the robot bounces on the rung at all, a tremendous impulse could tear the ¼-20 bolt out of its threads in the middle plate. This would ruin the main plate part because we didn’t think to leave space for a nut (the drive rails are directly behind). The big spacer mitigates this risk.
  • Removed the grayhill encoder because the winch neo’s built-in encoder gives us an elevator height resolution of about 0.06"

We designed a retainer plate to hold the pivoting arm steady while we are driving around. When the telescope rises and arm pivots back to prepare for the climb, the retainer plate will fall off and stay out of the way for the rest of the match.

One last thing—Onshape released an awesome update last week which could make CADing with box tube and 8020 so much easier. In the future we will try to integrate these new features into our CAD workflow.

1 Like

Ball Path

Our shooter team cut the poly carb for the floor of the chute, the standoffs, and the two fixed angled bars. They also tapped the standoffs, drilled through the chute bar to get the chute floor on V2, attached and centered the wheels to the shooter, put the poly cord onto the chute, and connected the shooter to the robot, finishing the V2 shooter!

Our chute team cut, faced, and milled all of the box tubings and put vinyl on some of their components. They also riveted the chute uprights to the shooter. While working on the hopper, they noticed that the distance between the pulleys of the shoot is too small, not providing enough ball compression. We fixed this by making the pulleys larger.

Our chutake (the hopper between the intake and the chute) has an issue where the polycord comes off the pulleys and jams. It only happens occasionally, but we cannot fix it by reversing the motors.

polycord_jam

high_shot polycord_jam2

We are considering several changes to mitigate this problem:

  • Taller vertical pulley walls
  • Replace polycord with wide polybelt
  • Add neo550-drive flex wheel and remove intermediate polycord (CAD screenshot below)

Unfortunately, this design has an interference with the intake polycord, and we haven’t thought of a simple solution yet.

Climber

We have been using a prototype climber and “final” climber to concurrently work on cargo and climbing. The image below and to the right is the final climber assembled to the robot. Partly because we transitioned from 8020 to box tubing and rivets, our fully assembled robot weighs only 90 pounds this year! That’s our lightest robot in recent history by 20 pounds.

The left image shows our 3D printed limit switch mount. This helps ensure that both telescoping arms are in sync with each other. We can also see the aluminum block which holds the telescope arms and rotating arms in place until we lower the telescope for the first time. This was necessary because our motors in brake mode were not strong enough to hold down the telescope (6:1 reduction with 0.5" thunderhex winch).

We are designing a different retaining mechanism that would hold the arms out of the way of the camera and shooter so that we can have a horizontal crossbeam connecting the two rotating arms. This would stiffen things and ensure that the rotating arms stay in sync.

square_climb

Here is our prototype climber in action. It’s assembled to last year’s offseason drivetrain. Although our climbing sequence is on the slower side, we do not swing at all and our hooks are deep. We hope to climb every match.

Inspiring Student Interest in Team Photography

For many years the LigerBots had trouble inspiring student interest in taking pictures of the team. We relied on our graphics and photography mentor to document almost every team activity herself. When the mentor stopped taking photos to concentrate on other aspects of team graphics, we had several years when very few photos were taken and our activities were not well recorded. We feel that this may also be a problem for many other FRC teams, and we would like to share how we have learned to crowdsource our team photography, increasing our student photography group from zero students to six regulars and many other occasional contributors.

  • We added a photography seminar for everyone on the team to our rounds of fall training. We got lucky that one of our student engineering leads was interested in photography. After he received some photography training from our mentor, we put him (rather than the mentor) center stage in front of his peers during the team training presentation. He demonstrated photographic concepts using DSLRs and phones. He got just technical enough in his explanations to show our engineering students how technologically complex and interesting photography can be, and just general enough for everyone on the team to see that they could easily take good photos themselves. At the end of each training session we had students practice taking pictures on their phones, using the advice we gave them. We found that having a formal training session like this helped legitimize photography as an important team skill, and made students feel confident that they could learn how quickly. Having a student who was already a respected leader do the training also helped increase its legitimacy with the rest of the team.
  • We ran informal five-minute seminars for new photographers. The student photography lead or the mentor gave new photographers who missed the fall training a five minute photography seminar, using whatever equipment the new student had. We gave the student a list of hints, sent them into our workspaces to take a few pictures, and gave them immediate feedback on how to improve the pictures. The main hints were about composition:
    • Take a lot of pictures so one of them is good.
    • Get hands, faces, and tools in every picture.
    • Move around a lot to get your shot. Get high, get low, get sideways.
    • Fill the frame. Don’t leave a lot of empty space at the top.
    • Don’t cut off important parts of your subject
  • We added a photography channel on Slack. This gave us a forum for discussing photography issues at all levels of interest, and further legitimized the subject as an important team skill. We pinned to this channel our shot lists, a Google slide show for photography training, event photography staffing lists, and instructions on how to transfer photos to the photography mentor so they could be put onto our team Flickr. (A future goal is to train a student in how to do this job.) Our mentor started calling out excellent photos on Slack as “photo of the week” and posting specific details about why the photos were good. This inspired pride in our team photographers, and fostered even better student photography. Our photo transfer procedure is:
    • Upload the photos to a specific folder on our Google Drive, with the photographer’s name on a sub-folder.
    • Ping the photography mentor on Slack to say this has been done.
    • The photography mentor downloads the photos from the Drive, color corrects them, edits them down, and puts them onto the team Flickr.
    • The mentor deletes them from the Drive to confirm with the photographer that the photos have been received.
  • We gave specific assignments to student photographers, with shot lists. We created a Google Doc with lists of the kinds of photos we needed for each important category of team activity: training, robot build, robot competition, outreach events, etc. This helped our students think like photojournalists, who go to an assignment with a list of pictures they must obtain to tell their story. Assigning students to specific events helped them feel ownership of the record for that event.
  • We made it clear that photography did not need to push aside a student’s other team activities, and that any equipment will be fine. One of our most important messages has been that our student photographers need to take pictures at every meeting, but only for a very short time. Five minutes will usually do it. This helps them feel that they don’t need to choose between their main interest on the team, (be it building the robot or soliciting sponsors) and team photography. We also made it clear that students did not need to own a DSLR. Phone photography is just fine.
  • We started a robot build blog on the Open Alliance. This blog provided a new incentive for our engineering students to document the robot build photographically as well as in words. A future goal is to get them as excited about posting on the team website’s general blog as they are about the Open Alliance.
  • We made team members more aware of our Flickr account as a resource for images for marketing, publicity, and for projects such as the Open Alliance build blog. Knowing that their photos would be seen by other FRC teams and the public also provided an incentive for students to take pictures.

The test of our new photography system will be whether it is sustainable once the current group of photographers has graduated. We hope to turn our new students into student mentors who will teach the next generation how to learn and enjoy LigerBots photography.

Photo of the week

We have been doing this since the start of build season, so we will roll them out once a day until we’ve caught up.

Why is this so good?

  • They thought to use a vertical format, anticipating that they wanted to catch the ball
  • They caught the ball in the air by using a fast shutter speed and finding the right instant to capture the peak of the action
  • Everyone in the group is in the picture (not cut off)
  • Everyone is looking at the ball, which focuses our attention as viewers. This is what elevates it from good to excellent.
  • Nothing is blocking our view of the action.
  • They’ve told us a story we can understand even without words: These people just shot a ball from the machinery that’s showing, and they’re watching it go up in order to figure something out as a group.
2 Likes