Has 6 wheeled designs surpased tank treads

[JVN]

Originally Posted by Pat Roche

Paul, the kinematical physics say that surface area does not matter.

Pat,
Re-read Paul's most recent post.
He is not disputing kinematics, just taking things to another level of complexity.

To be truly in touch with things, we must analyze the interactions between the wheels/tracks and carpet.

John

[Pat Roche]

Ok after reading Paul's second post (sorry amidst studying for finals is causing me great confussion) I've come up with this (This could be a repeat I'm not sure)

On a solid surface (i.e. concrete, etc.) a wheel and a track essentially have the same principle, the weight gravity force is "pulling" down on the bot and the surface is "pushing" back against the wheel/track that is in contact with the ground. The difference between a wheel and a track is the distribution of the wieght force of the robot. This doesn't matter however in the case of the hard floor because the way a wheel/track works is that static friction is applied in the same direction as whatever the wheel is attached to is moving. Friction does not include surface area in any way shape or form.

However on the other hand, while moving through sand or snow, a track has an advantage over a wheel because of the mass distribution. I guess the high heel and a snowshoe is a good analogy. I guess another would be a stick and a sand pail, but w/e. I pretty sure that there is math to prove that h/o i dont know it yet.

To be truly in touch with things, we must analyze the interactions between the wheels/tracks and carpet.

If I remember correctly from physics class, the interaction between the surface and the wheels/tracks is as follows.

No matter what the surface is made of because of the natural tendency of the weight force "pulling" down on the mass above there will always be the smallest depression. For an example of this I use a ball bearing and a sheet. If you hold a sheet taught and place a ball bearing on it the ball bearing will create a small depression.

Please correct me if I'm wrong or repetitive.

-Pat

[greencactus3]

ok i wouldnt say you are wrong, but missing a couple things. first, yes a snowshoe may be easier to walk over snow than high heel shoes, but high heel shoes if they actually puncture all the way through the snow into the ground, they will provide superior grip. seen in the high narrow wheels some serious mudgoing vehicles. so no you cannnot say tracks will always have an advantage over wheels in snow or sand. and if the ground was completely flat, yes what you are saying is completely correct, but if the ground was more not as smooth, wouldnt more in contact with the ground sometimes provide more grip? for example. velcro. more contact equals more grip. so well... i guess its not fair because it will be called "hooking into" the ground.

okay. im gonna stop because im just gonna make a fool of myself trying to disprove the laws of physics
so ill just correct you on the one thing im sure about.

Originally Posted by Pat Roche

If you hold a sheet taught and place a ball bearing on it the ball bearing will create a small depression.

Please correct me if I'm wrong or repetitive.

-Pat

that would be holding the sheet "taut" not "taught"

[Pat Roche]

Originally Posted by greencactus3

that would be holding the sheet "taut" not "taught"

Good catch

-Pat

[ShadowKnight]

Well, treads will actually provide a better traction system than wheels on very high torque low speed robot simply because they distribute the weight. Asuming coefficent of fricition is constant for any wheel/tread in contace with the ground, if the treads are distrubuting the pushing force on a greater surface area, the tread system will allow the drive train to provide greater overall force before the treads slip than a wheeled design. This is not because surface area affects friction, but because the force is distributed.

The decision to go with treads as opposed to wheels is depenent on the force exerted by the wheels/treads on the carpet. If in fact the force of the drive train will cause the wheels to slip, than a tread design may be preferred.

[Paul Copioli]

ShadowKnight,

The force being distributed along the tread has absolutely nothing to do with the ability to push, unless... the surface that your treads contact fails due to the contact. If the wheel design is wide enough such that the force imparted on the carpet does not make the carpet fail (i.e. tear apart like in 2002), then it will provide the exact same pushing force as a tread made of the same material provided the robot is the same weight.

Mike Norton's post is very accurate, but notice that the advantages of the tread design (disadvantages, too) really are not related to pushing force. Regarding Mike's point #3, I guarantee that if his team put a six wheel design with the same tread material they would still be in the top 5% of pushing robots.

Now, specifically regarding 2002: treads had a much better chance of not tearing the carpet than wheeled designs because they distributed the load. Tearing the carpet leads to less traction. The carpet of 2002 related to the possible weight of a fully loaded robot (around 500+ lbs) was like sand to a construction vehicle. Treads helped dissipate the pressure on the carpet to a point where the carpet would no longer fail. Wider wheels would provide similar functionality.

-Paul

[Manoel]

Originally Posted by Pat Roche

Paul, the kinematical physics say that surface area does not matter. A wheel and a track will slip once the coefficient of friction is broken regardless of surface area.
-Pat

Yes, but that is assuming ideal conditions. As others have said, you can have materials like Velcro, or rubber, that will "stick" to the surface. If you were to measure mu between rubber and asphalt, and used two blocks with the same weight but a different surface area, you'd get a slightly different result for mu (would it be linear with area?). Try it with Velcro and carpet and the difference would be huge.
For instance, extrapolate and loctite a block to a table. Loctite 158184 will hold 1300 N per square centimeter. Take a 1x1x1 (cm) block that weighs 100 g and pull it till you manage to rip it apart from the table. Would you consider your mu to be 1300 (Force/Normal)? Obviously not, because friction is not the only issue here. Same with the situations I discussed above.

Now I have two questions I'm not sure of the answer:
a) If surface area really doesn't matter, why is it much harder for a treaded robot to turn on a dime than it is for a robot with wheels? Also, why does it get much easier if you add an idler pulley? (I haven't put a lot of thought into this one yet, but I think it's not entirely dependent on the surface area, but also because the treads dig into the carpet and you have to overcome the fibers in order to skid laterally)

and

b) (not that it really makes much difference, but...) We all know that a rolling wheel has v=0 on the contacting point and v=2wr at the diametrically opposed point. Now, do we have static of dynamic friction? I would say static, considering if you "break traction" the force you apply to the floor will be reduced, and as the wheel won't be translating anymore, the relationship I mentioned above won't hold true.

[jimfortytwo]

I'm not prepared to say whether or not an increased surface area given the same normal force yields more traction. However, anyone who is arguing that it makes no difference whatsoever has both not actually tried it for themselves, and has not read the whitepaper on the subject.

[Paul Copioli]

jimFortytwo,

Was that aimed at me? If so, then I am ready to put my money where my mouth is. Are you?

-Paul

[Tom Bottiglieri]

I would trust what the UFH Engineer has to say about this.

Thats just me though.

[jimfortytwo]

Originally Posted by Paul Copioli

jimFortytwo,

Was that aimed at me? If so, then I am ready to put my money where my mouth is. Are you?

-Paul

No, I wasn't trying to single you out. Though I imagine you have not yet, in fact, read the white paper in question, and have not done the experiments yourself. Or are you holding out on us?

I suggest reading the whitepaper. Following your challenge, I reread it, and the results still somewhat surprise me. I won't bother summarizing, as it would spoil the surprise.

the paper:
http://www.chiefdelphi.com/forums/sh...threadid=15719

I guess the other part of "putting my money where my mouth is" should be to mention the traction tests I did for my team last year. We were looking at 8" hard rubber tires with an aggressive crosswise tread pattern. I did tests for kinetic friction with 30 pounds and 20 pounds of pressure on the tire. I was working with the risky approximation that in a 4-wheel drive robot each tire sees 30 pounds (30*4 =120), and in a 6-wheel design tires see 20 pounds (20*6 = 120). I can't remember the exact numbers, but my recollection is that I extrapolated a 10-20% benefit in the case of distributed load. Obviously a better procedure is needed to give any meaningful results, however.

I remain convinced that the traction advantage of treads over wheels is so highly dependent on the carpet and material that you can not generalize it. If I can put together materials to do a proper investigation of the subject, I will. Unfortunately we're on winter break, so that might have to wait until January.

[#1Transgirl1140]

My team has been around for 3 years this year and each year that we've been competing we've used treads. We've also used the same transmissions as well and hopefully we can use them again.