pic: 1918 NC Gears

This was our fall/pre-season project. We modified the “1918 standard” 8 axle drive base from our 2014 robot “Venus” to change it from a rectangle to an Octagon. We added another wheel at each of the mid positions. The new configuration is intended to improve performance against “T-Boning” or sidewall pinning. The original chassis was 28.75" wide x 26.5" long. The 2" side bulge added only 1" to the perimeter, but 4" to the width. We plan to field test it at the West Michigan Robotics Invitational (WMRI) on October 25.

I guess you’re pushing Truck Town…

… and the whole field.


And just when you thought 330 had a ton of versa wheels.

Out of curiosity, have you got any data on the effect of the stacked wheels vs. single wheels on frictional forces?

Also, is there any difference in the wear characteristics when you stack the wheels like that? I’d imagine it’d get expensive replacing all of those if they wore as fast as single versa-wheels.

That’s a lot of wheels you’ve got there. Make sure to save some weight for the robot!

You had plenty of wheels already, guys. Looking forward to seeing it in action later this month.

Looks great! One question though, do you drop the center wheels? And if so how much do you drop them? 0.100"?

Each wheel is supporting less weight so the wheels will wear slower individually (but you have to replace more wheels when they do wear out). The eventual cost of replacing wheels comes out to almost exactly the same. Pretend 4 wheels wear out after ~20 matches. That means that 8 wheels stacked in the same configuration will likely wear out after about ~40 matches. After 80 matches both drivetrains have used 16 total wheels.

Calculating frictional forces between non uniform, non smooth surfaces can be difficult because they don’t operate the same way that regular smooth surfaces do as is taught in basic newtonian physics. According to what you learned in those classes the surface area of an object shouldn’t change it’s frictional forces. However, in practice (you can test this yourself) having more rubber in contact with the floor can actually increase your drivetrain traction contradictory to what newtonian physics tells you. I’m sure someone on here can tell you how to calculate correct frictional forces between non smooth surfaces but the best way to find out is just to test it. Drive your robot against a wall and see how much current the motors draw before your wheels start to slip.

Yeah, I’m aware that modeling the frictional forces between these types of wheels and carpets is rather complicated. I was sort of hoping the OP had done some of those “robot against wall” tests as motivation for having that many wheels and could share the results with us.

So, in your experience, wheel wear scales pretty much linearly with weight supported? That’s good to know.

Geometrically, but yes, in my experience. It should pretty always, I can’t imagine why it wouldn’t, when a wheel rubs on a surface it loses an amount of material relative to hardness of the rubber, the hardness of the surface, and the pressure on the wheel. For a given pressure, the wheel will lose the same amount of material. This means that a wider wheel loses the same amount of material as a thinner wheel, the difference is that the thinner wheel loses depth much faster. If both wheels have the same tread depth then the thinner wheel will use up all of it’s tread sooner than they wider wheel simply because the wider wheel has more material on it. If the rubber and surface hardness are constant, the pressure should be the only thing to substantially affect wheel wear.

(reply to this and other posts)

We have never used VersaWheels individually - only as duals. 2014 was the first time we used them in any configuration, so we can’t really compare the performance and durability of “single” vs. “stacked” wheels. This robot has played about 100 matches. We did a complete wheel change after about 50 matches. The wheels on the mid axles wear differently than the ones on the corner axles. Mid wheels (which have .050" drop on this chassis) last longer and wear is uniform. The tips of the “W” treads become rounded off evenly along both edges of the tread tips. Corner wheels wear faster, and the tread tips take on a “sawtooth” appearance due to the rotation/scuffing that happens during turning. We rotated corner wheel sets left-for-right in between wheel changes to put the “sharp” edge on the “working” side of the tread.

We did some traction testing during 2014 build and measured about 200 lb of pushing force with new wheels (“sharp” treads) and a robot test weight of about 125 lb. We didn’t measure current draw. I agree that tread on carpet does not completely follow Newtonian rules of friction, but I don’t know what the actual physics is (are?). I don’t think the wear is directly proportional to wheel load either - the material interaction that improves traction also increases wear. In my judgment, the VersaWheels bite the carpet very well when new, but probably revert to “Newtonian” traction as the treads get rounded off.

VersaWheels are $5 each, so a full set of 20 is $100. I wouldn’t want to change them every day, but once or twice during the season is bearable. Right now, the treads are getting quite dull. They ran part of MSC, all of Championship, and the MARC. Only the new wheels are sharp.

There has been a lot of discussion on CD about the anti-T-Bone benefits of convex-sided robots (hex, octagon, round, etc.). I think bulged sides help, but I don’t think it is mainly because of the reduced friction/contact area between the robots (which seems to be the consensus among many people). A robot with “true radius” sides that is up against a wall or another bot can spin without moving (translating) its geometric center sideways. A straight side bot has to move its center sideways (and perhaps another robot, too) in order for the corners to clear the barrier as it rotates (a real challenge for long axis bots). “Venus Octobot” is about half way between radius and straight. We put more wheels in the bulge space, which increased the moment arm that powers the turn. We did some simulated T-Bone testing last week (with students pushing the 2011 bot into it) and our driver noticed an improvement in his ability to spin out of it. This Saturday, we get to test it at WMRI.

Thanks for the info! Have you ever done any pushing force tests with the same robot without the stacked wheels for comparison?

Nope. I wish we (or somebody else) did. I would like to see data on single vs. stacked wheels in the “new” and “worn” condition.

We field tested the “Octobase” at the WMRI. Overall, the event went very well for us. We seeded #1 and went on to win the event with the help of 2054 “Tech Vikes” and 2405 “Divided by Zero”. The modified chassis didn’t have any mechanical/reliability problems. We forgot to bring our team camera, so the only video we have is some brief snippets taken on phones - nothing that really shows off the strengths and weaknesses of the curved sides.

We really didn’t get into any extended T-Bone situations - just a lot of brief contacts. The curved sides seemed to do alright for open field engagements, but I think straight sides would have been just fine in those cases.

Our driver did NOT like the curved sides. His biggest complaint was the way they made the robot turn toward the wall (or another robot) whenever we made passing contact. He said it was like the wall just sucked the robot into it. We normally shoot from the corner, so this was a big annoyance. It required deliberate, careful action to avoid. Also, our driver also noted that he had to turn the opposite direction than before in order to get out of a T-Bone. He thought that this might not be a big problem for a new driver, but it really bugged him.

We are inclined to go back to a straight sided chassis next year. We are very glad that we tried this as a fall project. If we hadn’t done it now, we probably would have rationalized a reason to do it for the 2015 game.