Remember ball drive? It's back (in FTC)


#1

Remember ball drive, the TechnoKats’ drive base in 2003? Our team has been working on developing a new ball drive for use in FTC for the past few months, and we’re pretty proud of the end product. This one features only three motors and uses four custom-made rubber balls unlike the previous half omni wheel, half bowling ball configuration. Ball drive seems to be more suited to FTC because of our small motor limit and lack of reliable swerve drives.

Check out our reveal video here: https://www.youtube.com/watch?v=j09PzmVmzJw

Our website also has interactive CAD models and a list of answered FAQs. Check it out at http://balldrive.garfieldrobotics.com
http://data.garfieldrobotics.com/TransparentRender.png


#2

This is super cool! Probably the smoothest I’ve seen an FTC robot drive as well.


#3

You only know of 45’s cause I told you! :stuck_out_tongue:

After seeing your bot at WSR, i’ve been super impressed. This is no exception to what ive seen, and im blown away at this. Hopefully you guys use it in Rover Ruckus, and an event is out here so I can go check it out. Amazing job!


#4

Awesome job!

Side note - great job on the website, what did y’all use to make it?


#5

Not sure if this is what the above is asking, but I would like to know how you made the “custom-made” rubber balls, and if you have quantified what their CoF and compression is.


#6

I think Canon was asking about the website itself, which is very well designed.

As far as how the balls were made, this is from the FAQ on their website:

Here is a quick explanation. Our balls contain 3 inch foam cores that help keep weight and material costs low. We cast a ¼ inch layer of hard plastic around these cores, then cast an ⅛ inch layer of 70A hardness rubber to make a final ball with a 4 inch diameter. For more details on our manufacturing process feel free to contact us!

Edit: I forgot to say, very nice job!


#7

I’m in awe of this thing.

Are the omniwheels just overcoming the friction of the Colson on the ball when they strafe, or is there something else in play there?


#8

It’s probably not a huge consideration for FTC, but how durable have these balls been for you? Has it worn or torn at all?

Very cool design and implementation!


#9

That drive looks awesome!

However, I am wondering what benefits it offers over a “traditional” 4 omni holonomic drive. Especially considering the amount of space it takes up.


#10

Can you post some pics?


#11

I’m not them, but I can rattle off three quickly:

  1. Only three motors, compared to four. When FTC is rule-limited to 8, that’s not insignificant.

  2. Much simpler programming (check their website). Some teams just have that slant to them in the skills department.

  3. Better handling of terrain compared to many (surely not all, but many) omniwheel-based setups. Some mecanum-based setups will probably keep pace, but I struggle to see teams using omniwheels taking that slope with aplomb.


#12

This is too cool. You all did a great job, thanks for sharing it with us.


#13

For movement solely from the omni wheels, the balls would be rotating on an axis perpendicular to their contact with the colsons. The relative movement between the colson and the ball is probably fairly low (because it’s rotation rather than translation), leading to low losses. Ideally the resistance should be a pure moment, so it shouldn’t affect handling too much either*. I don’t think there’s any easy & better ways to handle this, so this implementation is probably about as good as it gets.

That being said, I have no experience with these drives, so I could be entirely wrong.

*this is the case for moving orthogonally, any non 90 degree angles could cause larger issues. But just like every other mechanical design issue, I’m sure the programming team can fix it.


#14

I’m curious how the control systems/ apparent handling / mechanical stability of the system would behave with different spacings between the ominiwheel pair on top of each ball. I am sure contact on the top of the ball leads to better efficiency, but you could really play with acceleration characteristics by changing those contact points.

woah, nice one on last year’s robot, finally a legitimate use for a scissor [lift].


#15

I hadn’t noticed that it’s only powered by 3 motors :ahh: I just finished picking my jaw up off the floor, just to have it fall back down again.


#16

The foam core wheels to me are genius definitely gonna have to try that in FRC for intake rollers and shooter wheels.


#17

Thanks for all the comments, I’ll try and answer all the questions that weren’t already answered by others.

Thanks, I just redid the site and I’m really happy with how it looks. We used Wordpress as our CMS and the Elementor plugin for page layouts with some custom parts as well. Elementor is super easy to use and I highly recommend it.

Essentially, yes. Matt G’s answer is correct - the balls are rotating on a perpendicular axis to the colson wheels so the friction is very low (basically none).

The balls have held up extremely well with almost no visible wear, even after 14 hours of driving on rough asphalt. Funnily enough considering how well they’re regarded, the colsons have shown the most wear out of any component on the robot. They’re basically flat now, but this is because they’re tightly tensioned with a spring against the sides of the balls and they’re made of softer rubber.


#18

How hard was it to keep the wheels spherical while adding the rubber? Seems like you would have to do a lot of fine layers.


#19

This design is pretty baller

Nice job!


#20

That’s the coolest thing I’ve seen in a long while. What a great application for a scissor lift, totally reminds of the first to I saw 148’s 2015 robot!