Re: Easy OmniTrack
 

The similarity between both our designs is amazing. Great minds think alike?

http://www.chiefdelphi.com/forums/sh...ad.php?t=46501

Re: Easy OmniTrack
 

Why not just set the two sets of tracks perpendicular to each other? It would be heavy but it would be able to push like a regular tank drive and maneuver like a holonomic.

Re: Easy OmniTrack
 

What is the advantage over traditional omni wheels?

This doesn't provide any more traction than those...

Re: Easy OmniTrack
 

Well wouldn't it provide more traction than omni wheels for the same reason tradtional treads provide more traction than regular wheels?

Re: Easy OmniTrack
 

Yes, it would -- if that reason existed. To a good first approximation, however, treads do not provide more traction than wheels.

Re: Easy OmniTrack
 

The way I always thought of making omni tracks simply threading a cable through a bunch of round rollers.
It's not a track exactly, but it's similar

Re: Easy OmniTrack
 

The issue then becomes on designing a wheel system that will drive it. Threads used toothes pulleys, but a cable would be harder. And yes, your idea is a lot simpler.

Re: Easy OmniTrack
 

Imagine a paddle-wheel with a V-shaped slot in the center of each paddle, with the V opening outward from the axis of the wheel. The cable would then be slung over the wheel so that the cable rests in the innermost point of the V. Each roller (they would probably have to have a diamond-shaped cross-section), with the cable threaded through them, would rest between paddles as the wheel turns.

Re: Easy OmniTrack
 

That was more or less the way I was planning on doing it

Re: Easy OmniTrack
 

Well I know from my basic physics class that surface area doesn't come into account when calculating friction force, but thats in an idealized environment, with no "chattering" and in which the surfaces are sliding past each other, not applying torque in the sense that we're talking about. But I thought that in the real world, surface area did play a role in how effectively a vehicle translated spinning wheels or treads into movement. If thats not true, then why do race cars have extra wide towers, and road bikes extra thin ones?

Re: Easy OmniTrack
 

read this post http://www.chiefdelphi.com/forums/sh...8&postcount=13

Re: Easy OmniTrack
 

Traction is a tricky subject. It seems to contradict itself all over the place.

In FIRST, the good brecoflex tracks can have a 1.5-1.6 coefficient of friction on the field (from a brecoflex catalog I saw a while back). The better tread a lot of teams use has a debated 1.2-1.3 cof (wedgetop/roughtop - variation comes from diferent materials and different test). I bet if the wheels had the same tread the belts did, they would have a 1.5-1.6 cof as well.

What I said above is fact, from experimentation (except the final guess). I still am not even close to understanding it completely.


On another note, these tracks would be better at climbing than traditional omnis.

Re: Easy OmniTrack
 

I'm confused on the whole traction thing. I've been shown the math on why surface area shouldn't influence traction, but here's why I disagree:

I built a kitbot with 4 andymark kit wheels, all powered from two BB gearboxes. It pushed a decent amount of stuff in our shop. Then I pulled the wheels off, and doubled up the andymarks. Now I had two wheels per shaft, effectively doubling the surface area on the ground. This bot pushed more, and could shove me across the ground when I was fully resisting. The other setup, with single wheels, could not shove me.

I have yet to take physics (that's next year for me), but I'm still basically familiar with the issue... Anyone care to hep me out here?

Re: Easy OmniTrack
 

Trust your experiment. Math does not control the world--it only attempts to describe it.

The friction equations in first year physics are basically only engineering approximations, and do not hold for sufficiently roughened surfaces, nor for surfaces of materials with unusual, non-linear properties--like rubber.

If you want to use the first year physics equations, then you can think of it like this--rubber has a non-linear "coefficient of friction." For most rubber material in most traction situations, the apparent (in other words, the experimental) coefficient of friction is higher when the rubber is more lightly loaded (less weight per area of the contact surface) than when it is more heavily loaded.

For a robot built to a maximum weight limit, more rubber surface on the ground means less pounds per square inch on the rubber surface, giving a higher apparent coefficient of friction and more grip for a given robot weight. So treads, or doubled wheels, have more grip than lower-surface-area alternatives using the same material.

Carlo Bertocchini, who built the combat robot Biohazard (and was on a winning FRC team if I remember right) answered this question several years ago (near the bottom of the linked page)

http://www.robotbooks.com/robot-materials.htm

Re: Easy OmniTrack
 

WOW! =O i've been working with this idea for a few years. Unfortunatly my CAD files were on a USB stick that was stolen at school. Heres what I have http://www.putfile.com/pic.php?img=6046594 it's in Solidworks(i'm not verry good at solidworks), i just installed Inventor 2008 the other day so i'll try and model it in there, it'll look alot nicer.

I think we need to discuss this, It's applications in FIRST are vast. And people are looking at torque more then menuverability, I think the instant directional changes will be more valuable then the ammount of torque the treads can put out, which happens to be only limited by the output of your gearbox =). Just so everyone knows, the treads should be placed in a Box formation, 1 on each side of the box, or in an H formation, this allows for multiple axis of stability and movement. I'd love to get into an AIM chatroom with you guys, my SN is Phillysteak527.