FRC 7461 Sushi Squad 2023 Build Thread

Welcome to our build thread!
There isn’t much here, but we had a lot we did these past two days.
You probably already read all the other OA posts detailing strategy, so I’ll just recap some of the main strategies and prototypes we did. Expect lots to come! (Or maybe not too much, most of the stuff will be on the discord, we will try our best, though!

Edit:

We also experimented with the drive train and the charging station today. We questioned whether the wheels would have enough traction on the polycarbonate for the robot to stay in place. Although the testing setup may have had issues that contributed to this, there was some minor slippage on a 15-degree incline. An easy way to mitigate this was to turn the wheels at an angle to improve traction, which held on very strongly. Although we aren’t sure about the sensitivity of the charging station as we haven’t built anything to test it yet, we imagine a potential tool we can develop is an algorithm to shift the position of the robot to utilize leverage to balance the scale even if alliance partners cannot accurately align themselves in the center. After overthinking the physics massively, we’ve begun coding a simple bang-bang-based solution and are looking to build a more robust solution once we have the resources to test it.

Goals:
Maintain driver uptime to as close to 100 percent as much as possible
Finish prototype bot by day 9-12
Finish Final bot by day 29

Prototype/Alpha bot specs:
L1/L2 Scoring
Shelf Pickup
A lot of drive practice
Potentially yeet cubes to l3
Reuse our old offseason chassis
Literally a 1dof arm on a pivot. It’s probably the same as the everybot

Final/comp bot specs/Priority List:


Comp Bot Priority list pt2 (assuming alpha derived):

  1. Floor cube pickup
  2. L3 Cube scoring
  3. L3 cone scoring
  4. L3 cone pickup (not very likely)
  5. L3 sideways cone pickup (probably won’t happen)

We also played with some auto-leveling code.

Some notes about auto leveling:
You should probably only have one auto-leveling robot at a time.
Figure out a way to stay on the platform after you level so that the other robot levels.
There are more posts on the way about this, I just wanted to get something out asap.
You can also stay still if the wheels are in an x shape on polycarbonate with basically zero slippage. In break mode, there’s a little bit, though.

Day 2 happenings:
Made two intake iterations:
Iteration 1 (did not record much):


Iteration 2:
Changed roller positions of bearings, hugging cone instead of gripping

(if videos don’t work, I’ll figure something out later)


Our freshmen were also shooting the cubes.

We popped the single cube like five times this meeting, do as you please with this information.

This is all alpha bot and super preliminary, so it will change fast. If you really want cad, I can send it to ya, though, but it won’t be very useful.

We want to finish by the end of this week/start of next week, but I need to check if my mentor is awake to order some stuff.

In theory, most of our alpha cad should be done soon. I’m getting tired, though… we will try our best!
Feel free to ask questions. Good luck with your season!

41 Likes

Could you clarify what you mean by driver “uptime”?

What drove the changes in iteration 2? From the video, it appears that iteration 1 held a lot firmer of a grip on the cone than iteration 2 (which seemed to struggle holding onto the cone).

The driver is driving as much as possible to

  1. Get as much drive practice as possible
  2. Have the robot be ran (and find problems) as fast as possible.
  3. Always have one operational robot or chassis at all times, and if there is free time, the driver can mess around on xrc sim. (or can also help with building robots)

Our goal is to build simple and fast, and with drive practice always being a crucial part of any game, we decided to emphasize it heavily this year.

We missed this, but iteration one has a very loose compression, which requires you to grip the bottom of the cone. We wanted more forgiveness, so iteration two is much tighter, and the center of the green wheel is in front of the center of the cone instead of in line.

That said, it could be better, and we need more testing to figure out things. Of course, iterations won’t always be better than previous ones, but we can always learn from testing.

Main tests we want to try out:

  1. Different wheel durometers to change some compression
  2. Adding a second layer to intaking to hold cone better
  3. Adding a star wheels for flavor

At the same time, we barely finished iteration two near the end of our meeting, so there’s a bit more testing that needs to be done.

(Please don’t copy these, they aren’t very good)
Iteration 1.

Iteration 2:


10 Likes

this is the definition of a grindset

3 Likes


Alpha bot cad has been rushed to a semi-complete state, note the incomplete intake. Design review + intake iterations needed
It’s just a 1 DOF arm with a claw that somewhat works, we need to iterate stuff and build it fast, but we hope to get it done soon.
Reusing our old drive base for alpha, so that’s why its 2x2, but we’ll run 2x1 swerve in comp.

Picking up cubes from a station seems like a hassle, so we might start prototyping and testing a ground intake that hands off to the grabber. (but only if it doesn’t hurt driver uptime compared to performance)

Have some orders that need to come in, and some adjustments need to be made (maintenance on the sprocket setup on the arm is difficult), but overall it’s getting places.

Anyway, good night.

12 Likes

how have you compensated for cones that are laying down?

1 Like

We’re just going to ignore it, aiming for just shelf pickup. There’s probably a way to do it. Waiting on another team or org to figure it out, but 99 percent most likely we’ll just not do it.

6 Likes

“Buddy that ain’t 5 degrees”

Legitimately caught me off guard and made me laugh hard as I have heard this exact type of response working in the real world so many times.

15 Likes

Loooooving how fast you’re iterating!
Thanks for sharing!

6 Likes

Slower meeting today, mostly am doing superstructure manufacturing, and tested a double roller inspired by vex tipping point.

Note by changing the speed of the rollers, or having 1 roller faster than the other, you can change the shot/release angle.

11 Likes

Preface

Today’s meeting was slow again, with some slight attendance issues. Hopefully, we can improve. We were able to get one iteration manipulator and some of our superstructure done. We are still looking to get done by day nine and on track for day 7 with prog time to spare (though of course, unexpected things will happen so who knows).

I’m really tired right now, I spent a lot of time trying to figure out geometry, so I won’t write much. I will answer questions tomorrow. Prog post coming soon.

Superstructure/Architecture


We built half of it, had some issues with max tube grid pattern and cutting at the wrong spot, 3d printing a jig, and writing a procedure to fix it.

Also had assembly issues, so we’ll fix them on Thursday. Mostly with figuring out what holes we were supposed to put gussets on. The goal is to finish the assembly on Thursday for the arm + superstructure and throw a claw on it. Need at least two days for basic code + electrical too, so hopefully, everything works out.

Manipulator

The testing isn’t very scientific. We’ll figure it out once we mount it onto our robot.

Design

I think I figured out a ground intake for cubes to the 1 DOF farm, it’s all in theory, though, just an over-the-bumper intake with some indexing wheels. I’ll post videos if it works.

We still want to keep the robot 1 DOF. The new manipulator primarily allows for a cone pickup from a single substation while keeping cycle times low.

It’s all theory crafting, though. It will be fun if it works by when we want to finish.

Anyway, design is a bit exhausting, so I can’t write anything much for now, but I’ll be able to answer questions tomorrow.

Other

Today’s been a long day, we will meet again on Thursday and post some progress, but mechanically not much is going on besides finishing up the robot and handing it off to electrical/prog.

Our field elements are happening too and having some measurement issues. (accidentally rounded up instead of down). I hope for the best there. Ideally, it gets done before the robot gets done, so we can get practicing.

Feel free to ask questions. I’m a bit unsure what to write, so questions are the best because someone else might think the same.

Unfortunately, nothing too innovative is happening, so apologies in advance.

Prog post coming soon!

18 Likes


i want to go to bed… (note, the robot doesn’t work - will fix issues tomorrow (today?) )

13 Likes

Great posts - but y’all need to invest in some safety glasses!

11 Likes

No meeting today!

Updates:
Claw was updated and fixed to fit packaging constraints as well as being an actual mechanism
Intake has pneumatics and bolts/bushings.
Tube heights have been raised.
Everything is finalized to begin assembly/MFG

Things to design:
Cone funnel
Cone shooter (?)

Looking to finish the alpha bot mechanically tomorrow, it’s a bit aggressive, but we’ll try our best.

Solo designer notes:
Sleep is very important, especially when you have to focus so hard on problem-solving in CAD. I had nine and a half hours of sleep one day and was extremely efficient when cadding, had five hours the next night and basically couldn’t do anything productive. Get your sleep even if it means not designing!

I think the hardest thing about this robot is keeping it as simple as possible while considering every single subteam and part of robotics to pull it off potentially (I don’t know if we can). There are lots of things to learn, and it’s still difficult to consider shipping lead times, what we have in stock, ease of assembly, ease of programming, etc. Lots to think about. This game was super duper fun to design for, and you can easily stay up late designing by accident.

Asking lots of people for help is also super useful to learn and design robots, thank you so much, Parker, for the design help today!

Anyway, no meeting today, so there are not too many details.

I still think you can hit a decent level with 1DOF, but it’s all theory crafting until next week. This game seems built on automation and driver practice, so we’re trying to go as fast as we can to make it work (we only meet four times a week officially :sweat: so there’s lots of planning to make the meetings work)

13 Likes

Day 6

Today’s meeting ran decently and could have been improved if we didn’t have so many shipping issues.

We finished manufacturing our entire alpha bot, and we got a bunch of assembly done, but we are waiting on things to arrive, so we aren’t entirely done. Mostly got our intake worked on today, as well as reinforcing our drivetrain for the new superstructure. The superstructure got repaired as well as we had some issues with cutting the box tubing wrong with maxtube grid pattern (i think mentioned in an earlier post. I printed a jig, but we’re always looking for more efficient ways, the tube is great though, it saves so much machining time. Tube plugs are awesome too.

The manipulator needs to still be built, but it’s simple enough that we can build it in the morning of our Saturday meeting, in parallel with finishing up our intake and arm.

With mcmaster somehow taking six days to ship from order, that kind of messed up our planning when the estimate was much sooner :sweat: We had similar issues with shipping last year, so I can’t really be surprised, with our kitbot drive train arriving week 4-5 last year lol. We’ve planned out accordingly this year, but there is only so much you can plan out and predict.

Currently, we are down 28t 20 dp gears, 6 inch Omni wheels, and a lot of tube plugs, so we need to find a solution to that tonight. I think a local team offered some 6 inch omnis, but we’re going to be putting in an order tonight for a bunch of stuff.

The piston we wanted arrives on Monday, so we have to wait till then if we want the robot to be truly “complete” I think there are lots we can do beforehand, and maybe we run string and surgical tubing as a temporary solution.

Can’t really get much prototyping done till the robot’s done, so we’ll probably have lots of videos once that happens.

Programming Accomplishments

Auto balance code working! Uses bang bang.

Also auto alignment with April tags works using odometry

Yeah, I don’t know how it works, expect a programming post soon though.

Mini motors mishap

We suddenly lost a few falcons, so now we can’t get a 2nd falcon drivetrain. I still havent gotten around to calling VEX about our four missing falcons from a year ago, but I think it’s too late now. I think we’re able to ethically source them from borrowing from a local team, so thank you to them!

Design

Mostly been doing refinements, so nothing new here, adding supports to the tube and getting reviewed for impact and loads.

Trying to get an absolute encoder enstalled 1-1 w/o the backlash, but it’s hard with the tradeoffs. Do I want mech to suffer or do I want to gamble on our assembly tolerances? Last time we did cross belt, we were too loose, and I’m worried that might be a problem, could add an idler maybe? Belt tensioning systems are hard if you want to make them perfect, but I might be overthinking it.

Design requires so much attention to detail :sob: it hurts my brain sometimes. Even 1dof arms are hard, this was wayyyy more difficult than I expected, and every design decision impacts the rest of the robot.

I’m having a new designer try to cad the cone shooter, though we need to get the field elements done before we can test and prototype it. We have a lot of room and weight avaliable, so I think it’s worth it to invest in this.

My internet is down today, and I dont have an adapter to attach my mouse to my laptop for a hotspot, so not much cad is happening tonight I think.

Maybe the CAD gods decided it was time for me to take a break. lol

Anyway, we have a large block of time on Saturday to get stuff done, hopefully, the robot gets mostly finished! Lots to learn in the incoming weeks, gotta fail fast and iterate fast, and hopefully, by building early we can get information from a real robot as soon as possible. We’re still pretty inexperienced, so there will be lots of stumbles along the way, but the experience has been really fun.

Still think this game is a software/drive practice game, so you got to have robust mechanisms that work consistently, and get lots of time for both on a working robot (your robot breaks a lot when you first build it).

Anyway, thanks for reading! Not sure what people really want us to post still, so questions are always appreciated. Trying to get more engagement so that we can help yall more.

Thank you!

26 Likes


edit: There is now a small superstructure, so the encoder setup should be easier, and the part count has been reduced by ~17 parts

The giant lever on the dead axel has been solved, which is nice.

The tradeoff is that single substation cones will be a lot harder if we want to do it, but I think I can figure out a solution by using 1339’s results in the future, for now, I’ll just ignore it because it makes building the robot much faster short term, as well as makes it simpler with fewer failure points, as well as makes the encoder setup a lot easier, in addition to decreasing the amount of weight at the top of the robot.

I’ll probably make the new encoder setup in the morning/evening, as we don’t have a meeting tomorrow.

8 Likes



Did not finish today :(, prob will finish robot around Thursday with basic autos

31 Likes

Casually shows up in the first week with an almost finished robot.

38 Likes

Avg david moment

5 Likes

This robot looks like its gonna win a week 1 regional.

9 Likes