Drivetrain First Thoughts?

[Jared Russell]

You can model a mecanum wheel as a regular wheel of the same diameter and tread material but with 71% the coefficient of friction (assuming 45 degree, friction-free rollers. In actuality, it is >71% for F/B and <71% for L/R due to friction.).

When you aren't slipping the wheel, you get the same speed out of a mecanum wheel as you would with a regular wheel of the same diameter. It is a common misconception (and one I had believed until recently) that a mecanum wheel is necessarily slower than any other wheel.

See here for the math: http://www.chiefdelphi.com/media/papers/2390

[Dr Theta]

Originally Posted by exprg:melonhead

I don't think we'll lose on speed but has anyone ever thought of acceleration being an issue with the omni drive? Just a thought since there isn't as much power in those wheels.

I believe we competed with you guys last year using a mecanum drive and anyone who saw it wouldn't say that it lacked for either acceleration or speed. If I remember correctly you guys do some pretty extensive scouting that involves taking pictures of every robot at competition I encourage you to look us up, also team 2052 from our area produced another very effective mecanum last year.

We actually will probably be going with a similar set up this year. If you want some help with anything feel free to PM me and we can probably provide you with some of the info you need.

[Madison]

Originally Posted by Jared341

You can model a mecanum wheel as a regular wheel of the same diameter and tread material but with 71% the coefficient of friction (assuming 45 degree, friction-free rollers. In actuality, it is >71% for F/B and <71% for L/R due to friction.).

When you aren't slipping the wheel, you get the same speed out of a mecanum wheel as you would with a regular wheel of the same diameter. It is a common misconception (and one I had believed until recently) that a mecanum wheel is necessarily slower than any other wheel.

See here for the math: http://www.chiefdelphi.com/media/papers/2390

Does anyone know, then, if the friction coefficients listed for AndyMark mecanum wheels are scaled in this manner? You wouldn't want to aim for 12 ft/sec and get 17 ft/sec because the values were already scaled.

[theprgramerdude]

Light cycle, anyone? Attach the arm mechanism to the center of the robot, in between wheels?

[Alan Anderson]

Originally Posted by Jared341

When you aren't slipping the wheel, you get the same speed out of a mecanum wheel as you would with a regular wheel of the same diameter. It is a common misconception (and one I had believed until recently) that a mecanum wheel is necessarily slower than any other wheel.

In theory, there is no difference between theory and practice. In practice, however...

Mecanum wheels' rollers do spin when driving in a straight line on carpet. This robs them of forward speed. Perhaps on a rigid surface like tile or concrete the "just like a traction wheel" model is adequate, but it falls short on a typical FRC field.

[davidcone315]

My team (3154) is seriously considering a mecanum drive, but I would like to know if the drive turns as well as moves horizontally.

[MrForbes]

If you build it and program it to, a mecanum wheeled robot can move forward, backward, sideways, at an angle, and also rotate about it's center (or any combination of these at once)

What it can't do quite as well as a robot with "normal" wheels, is get a lot of traction.

[Lil' Lavery]

Originally Posted by Ether

What do you mean by the word "inefficient" in this context?

By the very nature of the wheels, you're transmitting some of the force supplied by your motors into vectors other than your desired direction of travel. Because of the angeled rollers, you don't get 100% of the power of your motors at any given time (aside of rotating in place).

[Alan Anderson]

Originally Posted by Lil' Lavery

By the very nature of the wheels, you're transmitting some of the force supplied by your motors into vectors other than your desired direction of travel.

This is where someone will claim that those elseways-directed force vectors don't actually result in a loss of power because there's no motion in the undesired direction. That someone will be wrong, because the field surface is not perfectly rigid and does move sideways when the mecanum wheel tries to drive forward.

[Andrew Schreiber]

Originally Posted by Tom Line

Yes, most bots were long 6 wheel in '07. However, there was an added issue: to be a good robot you had to be able to ramp easily, and most ramps couldn't handle bots in a wide configuration.

Heh Heh Yeah...

Originally Posted by davidthefat

Have you guys realized that there were no height limitations during the game? that means technically if we can physically do it, we can make a 15 ft robot? Now what is the physical limit of the height before it become unstable? Now its some calculation with torque or something, but unfortunately I got an F on that test...

If you keep your arm light 15 feet is pretty simple to do. You can also deploy various outriggers to help keep you stable. But I have to ask, why would you consider this? You never need to go higher than 10'.

[Kevin Sevcik]

Originally Posted by Alan Anderson

This is where someone will claim that those elseways-directed force vectors don't actually result in a loss of power because there's no motion in the undesired direction. That someone will be wrong, because the field surface is not perfectly rigid and does move sideways when the mecanum wheel tries to drive forward.

So you mean "normal" wheels aren't 100% efficient either because the field surface shifts backwards when the wheel drives forwards, right? I'm not sure that makes much of a difference between the two wheels.

Do we even have a definition for "efficiency" in this case aside from the arbitrary demand that all force vectors should be pointed in the direction of travel because we think that's efficient? In that case, I'll point out that mecanums are vastly more efficient at turning than a 6WD, and thus are obviously superior by my idiosyncratic definition of efficiency.

[JosephBurns35]

i was considering omni drive but thought of the loss of power our team has done 4 wheel with rear omni's the last 3 years other than lunacy and have had a decent amount of success. but i was wondering how hard it would be to line up to the pegs with a non articulated arm.

[Ether]

Originally Posted by Kevin Sevcik

So you mean "normal" wheels aren't 100% efficient either because the field surface shifts backwards when the wheel drives forwards, right?

Yes, indeed it does. But it can probably be safely ignored for the purposes of FRC robotics.

In the case of mecanum wheels whose rollers have axial free play, the problem is exacerbated by the axial motion of the rollers. As each roller comes into contact with the carpet, it is forcibly pushed along its axle until the free play is taken up. As it does so, there will be some rotation. When the roller leaves contact with the floor, gravity and vibration re-open the gap. High quality (expensive) mecanum wheels with well-designed bearings in the rollers suffer less from this phenomenon.