4392 The Deceivers 2023 Open Alliance Build Thread

The Deceivers are excited to join the open alliance this year.

Last year we produced weekly videos recapping our accomplishments of the week. This year we have decided to expand and post more by joining the open alliance.

Links

More updates to come

11 Likes

Pre-Kickoff Update

We have done some updating to our swerve drive. Our swerve is based on the Coussens swerve but with some work to make it fit our needs better. We have moved to pushing the modules out to the corner, and updated to the REV through bore encoder. Last year we used the lamprey encoders to start but had issues with the zero getting lost at our first event. we switched to mag encoders for our future events and had very few issues with the encoders but were not quite happy with how they packaged. We still aren’t quite happy with the through bore encoders as they make the module assembly more difficult. We never got to testing the new lamprey encoders so we are not yet confident to use them this year. We hate dealing with tread so we also looked to possibly use the REV Swerve Wheel instead of a 3D printed wheel with tread. The wheel fits well and is what we will probably use if we can order some. Otherwise we will stick with out 3d printed wheel with tread.
Cad is here: Onshape

In the case of a game that has obstacles we will most likely not try to build a swerve to go over them. instead we will be using some form of tank drive. We have designed a basic 8 wheel 8 inch pneumatic tank drive to get us started if that is the case.
Cad is here: Onshape

We have done a lot of demos and presentations in the offseason as well. We have a video recapping all of that here: Rapid React, Off Season Recap; Team 4392 | The Deceivers - YouTube

2 Likes

Kickoff Impressions:

Open Field! We are gonna go ahead with swerve.
Cubes: They aren’t cubes, but spheres with points
Cones: Picking up from any orientation looks to be difficult. Single orientation is doable.
Charge Station: Seems to be one of the easier ramps to get up with the helpful second ramp. Building a viable simulation might be difficult for us.

We looked at many types of arms and lifts that might be suitable for this years game including a 4 bar, double jointed arm, Pink arm. We liked the simplicity of the four bar so we made a quick mock up out of 2x4s to see what it would look like. The dimensions were just for demonstration purposes and do not match robot rules.

4 foot extensions are scary. That is long way to reach. We are definitely going to want to keep the end of the arm light weight.

We do not meet sundays, so we will be creating our priority list monday, starting our prototyping, and starting to build our swerve modules.

3 Likes

Week One:

Aside from wheels which we are waiting on from REV, we got our drivetrain done. I think this is the fastest we have ever had a drivetrain built, and it is a custom swerve drive.

Initially we wanted to go with a four bar and wrist to score game pieces. That long of a four bar was scary and it didn’t allow for easy and safe pickup from the human player station as it would have to be extended when driving into the human player station. We definitely see us crashing with the arm out into the wall so we switched ideas. Intake with handoff and some sort of arm. We looked at both a Pink arm and a double jointed arm. The pink arm requires a lot of pieces and is quite complicated to build. A double jointed arm is fairly simple to build but complex in software. Luckily we have 3 mentors who have taken robotics classes at Lake Superior State University and are familiar with the controls for a robotic arm. With this info we decided a double jointed arm is the best route to go.

For our end effector our thoughts were couple of wheels and the ability to open and close. Opening and closing allows us the ability to quickly drop off game pieces and account for the size difference between cube and cone. The rollers will help with the handoff and give us the ability to softly launch game pieces onto the scoring platform.

When discussing our arm design changes, we discovered that actually, intaking standing up cones is the red herring. If you have the ability to score off either side of your robot by flipping your arm over (like a 2dof arm or pink arm) the orientation doesn’t matter as long as it comes point side first or flat side first. You just flip your arm over to flip the cone over and score off the other side of the robot. This led us to developing a roller intake that is touch it own it and then funnels the cone to the center where it can only be in one or two orientation. We built a prototype of this and really like it so we will continue the development if it.

We also have a video recap of the week

5 Likes

Quick update:

CAD is mostly complete. We need to finalize the intake and tweak our gripper a bit and then add the electronics. Otherwise it is pretty much done. Manufacturing parts is coming along nicely, and orders are starting to come in. Hoping to have the arm an drivetrain running by the end of the week.

I will be posting some more details about the robot as I find the time, but for now here is a weekly recap video.

https://www.youtube.com/watch?v=U-vrxZ8eN8s

5 Likes

Week 3

CAD is complete except for the intake.

The intake is long. Not a big fan of that. Our handoff also occurs outside the frame perimeter which also isn’t great. The plan is for the handoff to be completely automated, so maybe it won’t be that big of a deal.

Build of the rest of the robot is coming along nicely. Still waiting on parts to come in.

3 Likes

“This video contains content from SME, who has blocked it in your country on copyright grounds” :cry: :cry: :cry:

Do you happen to have a no music version that you can put up?

2 Likes

Excited to see it Chris! Is the upper elbow a virtual 4 bar or is it driven by a motor down low? I presume this can reach all scoring locations?

Here is one that was made using YouTube’s built in functions to remove copyright music. It’s a bit weird because it plays different songs where the original music was. It was Dolly Parton 9-5.

It is a virtual four bar. It is not a virtual four bar because we wanted the functions on virtual four bar, but rather that configuration allows us to place most of the weight down low. It does reach all scoring locations off both sides of the robot. I plan to write up a post on some details on the robot when I get some time. Hopefully I can do the arm soon.

1 Like

Here is a recap of week 4. Robot is coming along nicely. Still need to work on the intake, but there is a shelf we can pick up from if we don’t figure out the intake in time for week 1. Charged Up, Week 4 Recap; Team 4392 | The Deceivers - YouTube

I should have some time this week to write up some more detailed info finally.

2 Likes

Lets talk about resources

I love reading build blogs and looking at other robot designs, but I rarely come across a robot that I see and think we could copy it 1 for 1 because we don’t have the same resources. I figured I should make a post of our resources so other people can see how we operate.

Mentors

Our line up of mentors seems to be different than other teams

  • Our head coach works for the school and is a parent of an alumni. Their first exposure to FRC was 2012 when we were a rookie team, but she did not take over the head coach role until 2016.
  • I started to help mentor in 2012 when I was a freshman in college. My rookie year was 2008 on Enigma 2075
  • All of our other mentors are alumni from the team. Some are college students, some are not.

We basically have a total of 4 non college mentors that are at meeting frequently (at this point at least once a week), and 2 college mentors

This setup has positives and negatives. It’s convenient because everyone is familiar with how our team operates. On the other hand, we probably do some things that could be done better, but we don’t know any better. This is one thing I had hopped that The Open Alliance could help us find out.

Students

We come from a small school with a total of 144 kids in the high school. Our team currently consists of 11 students. We currently don’t have the resources (mainly mentors) to support an FTC team so we also allow some middle schoolers on the team.

This is fairly good size team for us. It has benefits for being easier to manage, but having less people available for things like scouting can lead to issues.

Build Space

We build out of a single classroom. The room is mainly used for us, but during the school day our head coach may have to use it. The front section with desks has to be cleaned up and ready to use as a classroom everyday but the back section allows us to store our robot and tools and field elements. While this is annoying, having a room we can keep everything is better than nothing.

Inside our build room we have a few main areas for work

3D Printing

Right at the entrance to the room we have our 3d printing area. It also stores the team laptops. We have a couple general purpose laptops and then 3 computer that are dedicated to CAD/Programming/Driver Station. We currently have 3 3D printers in our room. We have our Prusa Mk3s. This printer had an issue earlier this year where we needed to replace our hot end. While waiting for parts to come in, our sponsor, Bay Mills Community College (BMCC) offered to donate two Makerbot Replicator+s that they were no longer using. These are the printers we for most of our low load parts. BMCC also has two markforged 3d printers that we have access to.

We make extensive use of 3d printing parts in our designs because of these resources. Our belt pulleys, sensor mounts, and many other things are 3D printed.

CNC

In the corner of the room, we have an Omio CNC router. It is equipped with a mister from amazon, a tube magic jig from ozzyboards, and we use PVC as a waste board. Our mister uses a water filter housing as tank to hold the coolant (currently using cool mist), and air comes from a California Air Tools air compressor. We also store a vacuum under the router to clean up chips.

The Omio is our only form of CNC we have available to us. Basically every tube/plate on the robot goes through the Omio.

Other power tools

We have another table in the back of the room with our other power tools and small arbor press. We have a drill press, bandsaw, chop saw, sander, and mini lathe. The band saw is very rarely used. Besides the drill press and arbor press, the main tools on this table are the saw and mini lathe.

Saw

We have an Evolution Rage 3 saw. It’s a good saw for cutting aluminum tube. It has issues with cutting shaft. I have thought about upgrading to a band saw or something else. The clamp on the saw has worn out a bit and needs to be replaced soon.

Lathe

We have a mini lathe from Little Machine Shop. For the work we do and the space we have available it works fine. I would like to upgrade it with a DRO, but other wise it does what we need it to do.

We then have the normal assortment of hand tools: wrenches, t handles, drill bits, drills, rivet guns.

Meeting Times

We meet everyday after school from 3:30 until 6:30. We don’t require students to be at every meeting, just a set number of hours over the course of the season to attend competitions.

Sponsors

We have some amazing sponsors and community support. You can see a list of all of our sponsors on out website here

With their help, our program available at no cost to students. We cover travel/hotel so every student who wants to participate can.

2 Likes

my team is also making a two-bar, our’s is a climber in a box from andymark on a falcon gear box with pneumatic brakes and we are still experiencing a problem brought up in the video- the arm falling. what was the gear ratio you guys used to keep your arm up and stable with out tension or counter weight while driving?

We have roughly a 100:1 gear ratio. With the motor in brake mode it falls fairly slowly, but should be okay to hold for the 3 seconds between auto/teleop. During the match we use a simple motor feed forward that hold the arm up. In starting config, the arm will stay up even with no power to the motor. We will have some way to hold the arm upright while transporting the robot to the field.

I should also note that the arm isn’t super heavy. Total weight is probably around 30 pounds.

1 Like

Do you have the motors setup to brake mode and they still drop even at 100:1? What are you doing for a starting configulation?

They drop very slowly with 100:1 in brake mode. In starting config the arm is vertical and will stay upright even with no power. When moving the robot it can get bumped therefore we need something to keep it from falling on people.

ours is about 10 pounds and we are working with about a 1:28 gear box with added tension on the back and struggle with keeping it up but it makes sense that it only needs to be held for ~3 seconds when scoring. thanks for the help!

Been super busy, looks like I forgot to post week 5

https://www.youtube.com/watch?v=X_DDF0djDRo

Also here is week 6

https://www.youtube.com/watch?v=iS_Hj0p9AOs

3 Likes

Week 7 recap:

https://youtu.be/3yazOBV4XHw

Robot Teaser:

https://www.youtube.com/watch?v=qFdM5_dkeqE

2 Likes

Escanaba Recap:

https://www.youtube.com/watch?v=M9YT3Y_qKb4

Seeded 9th, Captain of 6th alliance, Quality Award. Overall a pretty good performance, but we were a bit upset with the amount of failures we had.

Murphy’s Law: What can go wrong, will go wrong.

In our first practice match, we had to get used to the insanely quick acceleration of the robot. We had been practicing but we don’t have carpet to test on. When we got on the real field we were much quicker than in practice. Practice match 2 was entertaining. We broke almost every single part of our robot. Starting the match we had run old auto code which didn’t have the fix we discovered in this thread. This caused us to run all the way across the field. In Teleop we ran past another robot that had their claw sticking out and ripped off our camera pole. The broken camera pole caused our intake chain to snap. Our sponsor panel was not attached rigidly enough and fell off. At the time we thought it was funny, lets get all the failure out of the way in practice. Sadly there was more to come.

Quals 7: we chose the wrong auto so we did not move and we missed our auto cone, Teleop started out great, but then we lost comms (we think we diagnosed it later) and then we bashed our gripper into the wall by the loading station. This broke our clamping belt and caused the end to fall off which meant we couldn’t pick up game pieces anymore. We went and balanced and wanted to try the drive off the side trick.

Quals 11: Picked up a cube off the floor but our position was off casing it to get stuck in our robot. Our partner got stuck so we tried to free them which also caused us to get stuck for a moment. We abandon that idea and went to get the cube unstuck and score more pieces. The belt skipped on the arm causing our automated positions to be off. We current use the built in PID of the spark max and a cancoder for initialization. We later added a rezero button but it was not functional in this match. We are thinking of changing this setup.

Quals 16: video is not available and don’t remember exactly what we did.

Quals 23: We more or less worked! we didn’t do cubes for a variety of reasons, and took it a bit slower, but the robot did what it was suppose to do.

Quals 28: Lost comms (again) in auto, the jerk after restoring comms throws off our arm again but we reset. Sadly the belt on our gripper failed again.

Quals 32: We did a systems check in the pit, right before this match and everything was functional. Go out to the field and the main breaker doesn’t turn on. No time to replace it so we are bypassed. After the match we opened it up and the plastic part the lever rotates had twisted and broke in half. We believe this was also the issue for our comms issue. That the lever never fully opened meaning bumps and hits could momentarily cause us to lose power.

Quals 43: Arm jerks in auto causing the robot to move and us to miss the cone in auto (we are looking for a fix to this). Arm position is off again and one of the treads on our drivetrain fails.

Quals 49: drivetrain encoder was unplgged (now hot glued in) during the process of changing wheel. Makes for a rough drive, but our driver handles it okay.

Quals 55: Gripper belt broke again, let’s score low/play defense

Quals 64: Starts off great, sponsor panel falls off and turns off out main breaker…

Quals 70: Arm is off again…seems to be getting worse. We diagnose the issue later.

Quals 74: finally a decent match, but our scoring positions are bit a low, we adjusted the pickup position to be higher which worked great.

Elims 4: tweaked all positions prior to going out there, we were ready. Everything worked, now we just need to be smoother about it.

Elims 8: Auto screwed up for some reason, but otherwise we were good, just get smoother.

Elims 10: Welcom back murphy. Our arm decides to go all the way around and we can’t reset it, our partners have issues as well.

Things we learned:

The issues with our arm came from our absolute encoder gear skipping or disengaging with the actual arm. this caused us to slowly over time lose our zero position.

We are changing our gripper belt from a spliced belt to a solid belt, hopefully that fixes the gripper failures.

We need to figure out what is causing the jerk in auto. It’s the same command in teleop and the return home is smooth.

We have a long list of improvements coming for LSSU and hope we can perform as well as we know we can.

5 Likes

Gripper:

The gripper consists of 8 sets of double stacked 2in thrifty squish wheels. A series of belts drives all the wheels and are powered by a neo 550 off a 3:1 ratio on each side.

We have another neo 550 in a 7:1 versaplanetary and a 2:1 gear reduction that drives the belt to close the gripper.

How the belt closes the gripper is with belt clamps. The left “finger” clamps to the top side of the belt, while the right “finger” clamps to the bottom side. This causes makes the belt close when driven one way and open when driven the other.

Inside the middle of the griper are a few sensors. We have a camera to aid the driver in picking up from the substation, a prox switch to detect and automatically close when something enters the gripper, and a color sensor which not only detects the type of game piece we have, but also when it is all the way in the back of the gripper.

On the arm behind the gripper are LEDs that allow us to signal to the human player what game piece we want next.

Besides the camera, all of the sensors and LEDs are controlled by a CANifier. The color sensor being I2C and the canifier not supporting any serial protocols was a bit of a challenge. We took this implementation of the raspberry pi pico color sensor program, but modified it to turn on IO that could be read by the CANifier. This saved us from having to run a bunch of wires down to the roborio.