FRC 1002 CircuitRunners | 2023 Offseason Robot Redesign

[EvanMeltz]

FRC 1002 Presents, Frank Castle, the successor of our 2023 competition robot, The Punisher

This robot blows The Punisher out of the water in every way.

This robot utilizes the same concept of our 2023 competition robot, of having an elevator at a fixed angle to extend out and up to the high grid, and adds our newly aquired MK4i Swerve Modules.

However, every mechanism has been completely reworked and heavy inspiration has been taken from many of our favorite robots from the world championship in Houston this season, including:

A roller intake very similar to 2910s, with a 3d printed pivot gear seen on FRC 125s Einstein Finals robot (the pulley + gear combo on the end of the pivot shaft is dead axle, there is a concealed bearing inside). All gears and pulleys are printed out of either a CF Nylon or CF Polycarbonate composite to reduce weight. Mecanums (assuming they were actually oriented correctly) vector the cone to the center of the intake for easier scoring, theoretically at least. No clue if that would actually work.

A 3-stage elevator, utilizing the WCP inline bearing blocks. Continuously rigged with polycord. Very similar design to 1678’s elevator this year, with an extra stage

A pivot arm for rotating the intake from the floor to scoring orientation. This entire mechanism is incredibly compact and made entirely from polycarbonate to reduce weight.

Lastly, the drivetrain. We’re using SDS MK4i’s with the same motor spacers that 2910 used this year to get L5 gear reductions. We’re also using the 90A TPU Tires from Formlabs and FRC Team 88

The electronics are located on the underside of the top cover, with a removable bellypan for easy access. The bellypan is held in place with 6x 10-32 rivnuts. The light blue blocks are placeholder ballast to increase weight and lower the center of gravity

Unfortunately, this entire robot is completely theoretical, and will probably never see the light of day. Its been more of a fun project over the summer to try and come up with the most dominant robot we can possibly think of for the past years game.

I’d be happy to answer any questions or take any feedback.

Here’s an onshape link: Onshape

[YYCDR]

Is the pivot arm going to be rigid enough? Just polycarbonate with no supports to help rigidity will make it super wiggly.

[EvanMeltz]

I’m not too sure, I was worried about this. I made it pretty thick, 3/8” polycarb, but I’d definitely want to stress test it before putting it in comp. There is room for improvement here

[David_IsMe]

we tried 3/8 poly carb and had to essentially bulid a box frame for it to sit in using 1x1 box tube.

[suPURDUEperAndy]

Love the design!

It might be just as light and more rigid to go 0.25" polycarb with some standoffs on the arm (or 1/8" aluminum with some standoffs, but then you need to worry about the arm taking a hit).

I hadn’t heard about the L5 spacers on 2910. It seems pretty aggressive. Got a weight estimate?

[EvanMeltz]

Without any ballast the robot is ridiculously light, only around 90lb with battery. I think that’s around the same as 2910, you can check their tech binder for more info about how their weight & gear ratio made them so insanely fast.

We originally planned on doing 3 module swerve like 4414 and maxing out the weight with ballast, but we couldn’t fit it with the elevator

[nuclearnerd]

I love the effort, but I also have some advice: I don’t think the wrist pivot as you’ve designed will survive. One good hit on a wall, grid or substation will bend the 1/2" hex as sketched:

(Here’s the math really quickly:

• A 100 lbf force on the intake plates (which feels entirely within the realm of possibility for a 100-150lbf robot) acting on a ~12" cantilever distance, will impart a 100 * 12 = 1200 in * lbf moment on the center of the hex
• Half inch hex has a section modulus Z of about 0.012 in^3 - You can measure that in OnShape
• The bending stress in a beam is M / Z = 1200lbf * in / 0.012in^3 = 100k psi
• The yield strength of 7075-T6 aluminum is only about 70k psi
• So your safety factor is 70k / 100k = 0.7, or less than one. That puppy is going to bend!
• Please correct me if I’ve made any mistakes above)

The hex carrying the herringbone pinion will also deflect, and may just skip teeth under load. Don’t trust long lengths of hex shaft like this!

I recommend you replace the wrist member with a 2x1 box beam with tube plugs in the end, or a length of Max Spline shaft., or even “Super Max Spline”. I also recommend you move the wrist motor to the arm, or add some extra plates so that the herringbone pinion is supported closer on either side so it won’t deflect.

[EvanMeltz]

Awesome technical breakdown on the wrist, thanks. I wish I knew how to do this kind of analysis. I guess that’s why I’m heading off to college.

Bending that hex shaft would be really bad. We oughta beef that up.

[ElliotS]

Look super cool, I am 100% going to check out the CAD when I get a chance. I am also wondering what is your frame dimensions?

[EvanMeltz]

27"x27"

[andej]

Some electrical/programming considerations I noticed:

1. Where do the sparkmaxes go? You’re using neos for both joints of the arm/wrist, and there’s no obvious place to put the controllers. Honestly, if you’re constrained by the number of Falcons you have, I would use falcons on the arm/wrist and neos for elevator drive.

2. I highly recommend zeroing switches/hall effects at the lower limit of the elevator. The integrated BLDC encoders are fine once you have a reference-- it can be easy to accidentally power on the robot while the elevator is in an unexpected state, and a software zeroing routine mitigates this.

3. Absolute position sensors on rotating joints for the same reason (cancoder, rev through bore, canandcoder all work fine)

4. You have plenty of upper bellypan space to use for electronics. This will make your life SO MUCH easier compared with having to flip the robot on its side and unfasten a big machined part just to do a routine repair or check.

5. Even if you ignore the previous suggestion, the roborio and radio absolutely have to move to a more accessible location. Radio preferably at the top of the tower. Your electrical people and programmers will thank you.

[EvanMeltz]

NEOs were due to size constraints at one point. Looking at it now I’m not too sure why they aren’t Falcons.

Having electronics hidden under the robot looks really cool in CAD but that does seem like it would be a nightmare.

Adding a zeroing limit switch to the elevator seems easy enough. I think a full redesign on the pivot arm + wrist is in order, it needs to be way beefier & I’d like it to change how the pivot is driven, so might as well redo it all with encoders in mind from the get-go.

[EvanMeltz]

something like that

[EvanMeltz]

Hows that look?

[asid61]

As an alternative to limit switches, a simple zeroing routine where you drive the elevator downwards and look for a current spike is fairly easy to implement. Then, override any elevator command (except manual control) with the zeroing one if something is commanded after a reboot.

[EvanMeltz]

that took a while