Treaded tank drive vs. Traction wheels

Coefficient of friction changes in that case. See my other comment.

Well, you have this past year’s robot, right? That would cut down the cost of testing some nice traction wheels significantly - you’re really just replacing the KoP wheels with performance wheels with some roughtop tread on them. That’s significantly cheaper than buying (for example) a pair of Rhino tracks, plus gearboxes and motors and a chassis to attach them to!

So, see if there’s any way to leverage your existing robot to help decrease costs and make your testing easier or more valuable!

The coefficient of friction is just an empirically derived constant that describes observed behavior. The dynamic forces at play are NOT as simple as “the coefficient of friction changes”, though - that would imply the tractive boost would be the same relative boost regardless of the normal force applied. This isn’t true in these dynamic situations - the whole linear relationship between normal force and tractive force can and often does go out the window.

This seems like splitting hairs, but you can’t just say “oh, roughtop that interlocks with the carpet is 1.3 and roughtop when it doesn’t is 1.1” or something like that.

This past season we modded the KoP frame to fit 6" mecanums. We used one of the Toughbox Minis for climbing, and had a Toughbox Micro for each mecanum wheel. After hearing all the responses in this thread… and considering we are looking into octocanum… I think it would be best if we went simple and reliable with traction wheels :slight_smile:

If one simply defines the coefficient of friction to be the (not necessarily constant) ratio between the friction force and the normal force, the claim is just fine - it merely becomes a rephrasing. Whether or not it is useful to frame the problem in this way when the linear model breaks down is, of course, another matter entirely.

Simply not accurate in FRC-world, as others have noted.

Materials *always *interlock on some scale because no surface is *perfectly *smooth. See this experiment with solid PTFE and felt that shows a clear dependence on normal force:

(Yeah, I know it’s not a peer-reviewed publication, but it makes the point)

About the closest to a ‘polymer on carpet’ model with lots of peer review I know of is racecar tires. I think that compliant tire polymers on asphault is a reasonable approximation of rigid wheel polymers on compliant carpet in terms of grip behavior vs load.


Specifically Figure 7:

Coefficient of friction is gained as load is reduced, but it’s not a tremendous effect. Halving the normal load saw a ~20% bump in grip (this not a linear relationship though!)

The issue we had their is we couldn’t keep tension on our tracks. When we got home and took our bumpers off we found that we bent our steel brackets for tensioning. Other than that we had no issues with tracks the entire time.

Both this year and last year, there was not a single robot we went against we couldn’t push. Our defense capabilities this off season is something else. We slam into them at 17fps then downshift to 8fps and push them where ever we want them

These were the tracks on our last years robot but the ones we use now are almost the same:

To the OP, ask the folks who want to use tank treads, how many elite teams use tank treads?

I can’t tell you how many times I have asked this…

Why would this happen for wheels, and not for treads?

Having been in FRC for many years now (and having once been a student dead-set on certain designs without regard for their actual suitability, myself), this sounds suspiciously like the kind of notion that gets volunteered when someone is searching for justifications for a position they’ve already decided to take.

What do you think about wheels with tread riveted to them?

In this case, you explicitly have “wheels of the same material as the tread,” so the treads likely have slightly more traction.

I’m not such a fan of these sorts of wheels, myself, because I find that none of the tread attachment schemes are satisfactory (cutting the tread to size and drilling it is a royal pain), and replacing it once it starts to wear is, likewise, a giant pain.

If you really want super traction-y wheels, what you want to use are (as I mentioned earlier) W-treaded VersaWheels from VexPro. They wear quickly, though, and so you have to replace them reasonably often to maintain the edge. (I suppose if you really wanted unparalleled traction, the solution would be something like the super-soft banebots wheels, so long as you’re okay with replacing them after literally every match and dealing with the irate field staff as you leave what amount to giant eraser shavings all over the field…).

We use Colsons, because one set will last an entire competition season, and find them to be reasonably traction-y (we might experiment with cutting patterns in them, as it appears to yield a traction boost without having any significant cost).

I don’t know why it would happen for wheels, and not for tank treads. I would guess that their argument was that more surface area resulted in more stability, thus a lower likelihood of causing it to happen for tank treads. I have no idea if this is right, or even if that was the reason for their “skipping” argument.

The Colsons look like a solid option, thanks for the recommendation. I remember seeing them on a lot of robots at worlds…

Let me ask it another way. What’s the goal of your off season project? There’s a lot of interesting talk about coefficient of friction versus true traction forces. Is your goal to study the science, develop a “go-to” drive train, or drive and play better?

Early in our team’s history we spent time on developing a “go-to” drive train to speed up design and construction during the build season. That’s definitely been helpful.

Recently, we focused on additional drive team practice and drive train software as a way to improve. Driver interface and the associated software have really made the biggest impact. Last thing I want to worry about is the drive train hardware.

Good luck with your project.


If you have observed this behavior than you can…

Regardless, I was merely pointing out that buddy changed the context of my comment to make it wrong. That or assumptions were made. I was stating that neither drive train has an advantage if the coefficients of friction are the same.

Our goal is to create a “go-to” drivetrain like you mentioned which is capable of speed/maneuverability, playing defense, and using motion profiles in auton. Our driver really liked the translation capability of mecanum, so I think after this thread, we’ve decided to experiment with octocanum.

I can agree with the treads breaking. We had many issues last year with our Rhino Tracks last year, and they seemed to break at the worst moments (Playoffs @ Ann Arbor Skyline 2016, WMRI Playoffs, etc.) Once they break, you are done. If used in a match, wheels are more reliable since if you lose one wheel (somehow), you can still move. We went with a West Coast Drive this year, which allowed us to keep going unless something happened to the gearbox itself, a motor, or the center wheels. We actually drove a practice WCD we had with only the 2 center wheels and we could move. It drove HORRIBLY but if everything started to go wrong, less we have to rely on.

It’s not the best video but if you watch at the top of the screen team 5464 has full Rino Track and they pushed around all the teams and never got pushed back. Best example I can think of.


This is a great question and I’ll be pleased to see the results if someone gets to do some instrumented tests with identical vehicles.

If you think about treads used in the real world, they’re used on tanks, on skid-loaders and they’re common on European tractors. Being used more on tractors in the USA lately. Mostly there for a couple reasons, but the foremost is to support a heavy vehicle on a soft surface.

This could occur in a robotics match, but to my knowledge, we haven’t had a real ‘soft’ surface yet.

Treads also help to bridge a gap over rough terrain, but with large enough wheels, that wasn’t necessary in Stronghold.

Treads also provide an advantage in that you can drive them from one drive wheel where if you had six or eight regular wheels, you’ll need chains or belts or something between them.

Treads are always bad to steer, using more energy, tearing up the surface, but wheels can steer in multiple ways. Large military vehicles even have multiple-ackerman steered wheeled axles.

Treads are very resistant to pushing forward and back, but I remember seeing treads pushed off of their wheels sideways in Stronghold. I don’t recall seeing a four-treaded H-drive or Octonum. That would be interesting, but expensive and complicated.

I think wheels will always have a higher top speed than treads. Feet are faster than snakes most of the time too.

If this is true I recommend tank drive 100%. Octocanum is great if it works well, and there’s nothing wrong with researching it right now, but make sure that you’ve got a good tank drive down to start and maybe make having a good Octocanum your goal for the 2019 season.

I’m guessing you, like me, weren’t around in 1992. I suspect treads could have done well in that game!

Had to look that one up…yes, I think treads might have been good.