Colsons vs. VersaWheels vs. Treaded Traction Wheels

[Osseus_Dominum]

It doesn't actually say on the product info pages on VexPro, but I suppose it's a safe assumption that the coefficient of friction is tested on FRC spec carpet. So given that all of the CoF values for these wheels are very similar (or identical), which is the best overall for traction? It seems to me that the VersaWheels and the Tread on the traction wheels should give more grip than a Colson with the full weight of a robot because the treads can dig into the carpet unlike the smooth surface of the Colson rubber. Since we don't know how these wheels were tested, it's possible that this wasn't even a factor when measuring the CoF because they may not have loaded the wheels enough. So which style of wheel (and tread) really offers more grip?

On a related note, I haven't seen anyone discuss wheel size when talking about traction. It seems to me that a bigger wheel should give more traction since there is a larger area of rubber contacting the carpet. What do you think?

[Andrew_L]

Having used all three kinds of wheels, here are my experiences:

Colsons: I haven't found tangible evidence that a wider colson wheel provides more traction than a thinner one, but this is mostly due to lack of serious testing. The beauty of the colson wheel lies in the soft rubber's ability to heat up during a match and slightly deform such that it "digs" into the carpet in the way a treaded wheel does. Another advantage is how incredible colsons wear, to the point that you can run the same pair all season and never need to change them. This cannot be said about the other kinds of wheels.

VersaWheels (W-tread): Personally, I think the W-tread versawheels have the most traction out of any wheel listed here, with its hard plastic cleats clawing into the carpet when run for the first time (and in my early tests, clawing into my skin with ease, requiring a quick visit to the first aid kit). The problem with these wheels is how incredibly quickly they wear down. As soon as 3 matches in a fresh set of wheels will show signs of rounding on the normally pointy jagged cleats. Once some rounding appears the effective life of the wheels is nearly over, with a few more matches of hard driving they'll be worn down to near round with some small bumps, losing the grip and traction from the now worn down cleats. The key to using these wheels successfully is to swap them out every few matches. I personally haven't swapped out more than twice an event, but I have friends who drove their robots more aggressively than my team who had to go through 3-4 sets of wheels per event. These are great wheels if you have enough stock to swap them out multiple times per event. Just keep in mind that on a 6 wheel drive robot, 3 swaps per event through 3 events is 54 wheels, and at $6 a pop, that's $324 you'd be shelling out just to keep good wheels on your robot through the season.

Treaded wheels: Treaded wheels are always a solid option if you choose the right tread. There are many different treads out there to use, and I know some people have their tried and true special tread they like using (cough cough green 469 tread), but from my experiences there are really only two treads worth considering: Blue nitrile roughtop tread, or tan wedgetop tread. Blue nitrile roughtop tread is the ideal FRC robot on carpet tread. They last long enough that you don't need to swap them out nearly as much as you do the versawheels (but you should be swapping the tread out during season) and they provide enough traction to get through most pushing matches. The tan wedgetop is a niche tread that shows its strengths with non-carpet materials. This tread style was popular in 2012 when there were lots of non-carpet surfaces for the robot to drive over and interact with. On these surfaces, the wedgetop tread had greater traction than the roughtop treads did. It has also been found that colsons also have more traction than most treads on non-carpeted surfaces, though I have not seen a test that compared colsons and wedgetop tread on non-carpet surfaces to see which ultimately had the most traction. Either, however, holds more traction than roughtop tread on these surfaces.

With versawheels and treaded wheels, I have seen evidence that a wider wheel provides greater traction because there is more interlocking surface area with the carpet. Logically, this is not relevant on non-carpeted surfaces that the wheels cannot "dig" into.

These are just my experiences with each kind of wheel, as always, your mileage may vary. Personally, I intend to use colsons on most robots I design in the coming future because the ease of maintenance (aka the fact that there is none) and the universally competitive traction on nearly all surfaces one would encounter on an frc field.

[Cothron Theiss]

Tread is grippier than VersaWheel is grippier than Colson. Which leads directly to Colson being more durable than VersaWheel which is more durable than Tread. In my opinion, and from what I've seen, the difference between Colsons and VersaWheels is greater than the difference between VersaWheels and tread. Also from what I've seen, the white KOP wheels perform very similarly to VersaWheels. I have no data to back those statements up, but that's what I've picked up on.

About bigger wheels providing more traction: the diameter of the wheel has no noticeable effect on traction, but the width absolutely does. So doubling up on traction wheels to create 2" wide wheels will theoretically double your traction. Will you actually get twice the pushing power? I doubt it. But your traction will increase significantly.

EDIT - Andrew's post reminded me that W pattern VersaWheels are a thing. I've never used them. I don't think I've ever seen them in person. So everytime I said VersaWheels, I meant specifically the Diamond tread.

[Lil' Lavery]

Speaking anecdotally, W-Tread VersaWheels are the highest traction wheels I've experienced in FRC in the lateral direction. If your goal is to resist turning moments and/or being pushed sideways, you're not going to find a higher performing wheel. The points in the W-Tread will essentially act like cleats when pushed sideways. (edit: watched some old video, not quite as good in the lateral direction as I remembered, but still an overall very high traction wheel. If there's a flat floor game that facilitates 4" wheels next year, we will strongly consider W-treads)

In the transverse direction, the difference is less pronounced, and there are several wheel/tread options that will get you into the same ballpark.

In terms of drawbacks of the W-Tread VersaWheels, I'll echo Andrew that the biggest one is tread wear. Over the course of an event, you will run the points off the wheels until they begin to approximate round. In fact, when we added a high goal shooter to our 2016 robot during the post-season, we used a set of rounded-off VersaWheels we had taken off our 2013 or 2014 robot (as they were now so round they could be used as shooter wheels). This tread wear will not only impact your traction, but also noticeably affect the effective wheel radius. As the tread wears, the wheels get smaller. We actually lost a quarter-final match at MAR Championship in 2014 because our autonomous stopped slightly short of where we expected it to, causing us to fail to score. Due to this tread wear, I would at recommend at least having one spare set of W-Tread wheels ready at each event. You can either swap them out overnight or before your elimination matches (or more frequently if you can afford it). You can also get a bit more mileage out of the treads if you rotate your higher and lower wearing wheels (our 2014 sat mostly on the
"back 4" wheels, causing them to wear faster than our front wheels). I don't think you'll need quite as many wheels as Andrew suggested (especially if you re-use your elimination wheels from one event to start your next event), but the costs of replacements can add up.

For comparison, the Colson (or NPC) wheeled robots I've built have not needed wheel replacements, and blue nitrile roughtop treads on traction wheels have lasted nearly two events before significant wear appeared (granted that robot was a good 15-20 lbs under the weight limit).


Additionally, because of the very high lateral traction, you'll want to be conscious of your wheel placement, wheel drop, and loading. Turning scrub can be a serious concern for any high-traction drive, but the higher your lateral traction, the bigger the issue can be if you don't plan for it. Excessive scrub can lead to significant current draws during turns, potentially risking brown outs (or even breaker trips in 6 CIM drivetrains). It's an issue that can easily be worked around, but you have to be conscious of it.

[mman1506]

Originally Posted by Cothron Theiss

About bigger wheels providing more traction: the diameter of the wheel has no noticeable effect on traction, but the width absolutely does. So doubling up on traction wheels to create 2" wide wheels will theoretically double your traction. Will you actually get twice the pushing power? I doubt it. But your traction will increase significantly.

Doubling your wheels surface area does not double your traction theoretical or not. The only way to increase traction is to increase the weight of your robot or raise COF of your wheels to the carpet. I suggest reading this document http://www.simbotics.org/files/pdf/drivetraindesign.pdf

[ThaddeusMaximus]

Originally Posted by Andrew_L

Having used all three kinds of wheels, here are my experiences:

Colsons:The beauty of the colson wheel lies in the soft rubber's ability to heat up during a match and slightly deform such that it "digs" into the carpet in the way a treaded wheel does.

Now I want to try out heat-gunning some Colsons pre-match...

[IndySam]

Originally Posted by ThaddeusMaximus

Now I want to try out heat-gunning some Colsons pre-match...

Or make little tire warmers like F1 uses.

[BumblingBuilder]

On King TeC we like to stick to Blue Nitrile tread on our wheels, in which we usually try to stick with 4" performance wheels that we mount them on.

They are much more durable than any other tread, as well as your Versawheel tire. While I have not used the versawheel tires personally, I do not really like the wheels themselves (we riveted blue nitrile instead of buying the tire tread when we tested them). We hated having to put the wheels together, just to have, at the end of everything, the hubs somehow out of alignment just enough for VexPro's hex shaft to not fit easily (which is one thing we value in wheels: being able to remove them in a pinch). We ended up fidgeting with 6 wheels for a long time just to get it to fit well. We wasted a lot of time on these wheels, and we didn't even use them on our robot. I was against buying them from the get-go, but that was against the desires of the leadership.
I really want the time I wasted that day back.

Colsons are just too heavy for my liking and don't grip into the carpet. Blue nitrile doesn't wear quick enough to make me irritated with having to frequently retread, so it is naturally my choice when pinned up against colsons. However, you can carve patterns into the rubber to grip the carpet more, but I would just rather not deal with doing that when I can have a lighter blue nitrile treaded performance wheel.

I guess it all comes down to circumstance, preferance, and what you are willing to do. If you need more weight on bottom to balance out the center of mass, colsons won't hurt. If the cleat-like attribute of versawheels are particularly alluring and you don't mind having to assemble the wheels and replace tread frequently, go for it. If you enjoy the smell of blue nitrile (yes it smells really good), satisfy your nose.

On a side note, has any team tested pebble top tread yet? We want to order it on King TeC and test it one afternoon, but if a team has tested it and found it undesirable, well save our $32.

[RyanCahoon]

Originally Posted by BumblingBuilder

While I have not used the versawheel tires personally, I do not really like the wheels themselves (we riveted blue nitrile instead of buying the tire tread when we tested them). We hated having to put the wheels together, just to have, at the end of everything, the hubs somehow out of alignment just enough for VexPro's hex shaft to not fit easily

Those sound like the VexPro Traction Wheels, rather than the VersaWheels, which come with the tread molded onto the wheel hubs

[EricH]

Originally Posted by mman1506

Doubling your wheels surface area does not double your traction theoretical or not. The only way to increase traction is to increase the weight of your robot or raise COF of your wheels to the carpet. I suggest reading this document http://www.simbotics.org/files/pdf/drivetraindesign.pdf

There is some debate as to whether increasing surface area in FRC applications will increase traction, with anecdotal evidence that more surface area does in fact increase traction. I believe someone was going to run a test, but I can't recall if they ever published results. Would be an interesting experiment to design, run, publish, and see if someone else can replicate it.

The basic premise is that widening the treads on the wheels--of all types--provides more room for wheel tread to interact with carpet. The interaction, at a micro level, isn't all that different than an athlete's cleats digging into the turf; wider wheels mean more places to put a spike on the cleats. The effect is seen as more traction to the robot. (If you really want to be technical, it's more of an increase in CoF than anything else--but it's seen as more traction.)

Now, it won't double the traction force--but you might get a pound or two across all however many wheels you've got.

[BumblingBuilder]

Originally Posted by RyanCahoon

Those sound like the VexPro Traction Wheels, rather than the VersaWheels, which come with the tread molded onto the wheel hubs

Ah. My mistake. Thanks for clearing that up.

[T^2]

Originally Posted by mman1506

The only way to increase traction is to increase the weight of your robot or raise COF of your wheels to the carpet.

Oh hey, this fiction again. I wonder why F1 cars don't use small skinny tires to save weight?



That aside, traction is far from the end-all of wheel selection. Years ago, when most teams had bad drivetrains, traction was pretty important for winning pushing matches when playing defense or counter-defense. More recently, though, it seems subjectively that high-level play has shifted more to zone and tempo defense. Offensive robots are also more inclined to evade rather than push through obstacles. Most off-the-shelf options, Colsons included, are adequate for field traversal and zone control, which means that ease-of-use takes higher priority.

[RoboChair]

Originally Posted by mman1506

Doubling your wheels surface area does not double your traction theoretical or not. The only way to increase traction is to increase the weight of your robot or raise COF of your wheels to the carpet. I suggest reading this document http://www.simbotics.org/files/pdf/drivetraindesign.pdf

If you are using textured tread, you will get more traction with the carpet using a wider wheel. The effect is not huge compared to changing robot weight, but it is very real. Also all friction materials have a maximum power transference capacity for a given surface area, if you exceed that you will burn rubber and prematurely wear your wheels. A wider wheel will wear slower which keeps more tread longer, which means more overall traction.

[jwfoss]

Ah yes, the great traction wheel debate.
Has anyone done any actual testing with direct comparison recently? I'd love to know more about the test rig that was used.

In my experience, width, diameter, and material all have an impact on traction. We however have standardized onto Colson's because of their wear and grip properties on the variety of surfaces that FIRST likes to throw at us.

[FrankJ]

Originally Posted by T^2

Oh hey, this fiction again. I wonder why F1 cars don't use small skinny tires to save weight?

F1 cars have considerably more power than an FRC robot as well as multiple G loading. Anyway. Point taken that the classic friction model (F=COF*N) really only applies with relatively smooth and inelastic objects.