Re: How Do mecanum wheels handle the bumps?
 

Your description makes sense: for each wheel, add the force from the other wheel as well as its axle and each wheel ends up with zero sideways force and full forward force. To jump into the case where the robot is moving, your description seems equally valid (with the same assumption that the rollers stick well on the floor).

If you look at a mechanum bot going at 45 degrees, you'll see two wheels on opposite corners don't turn at all, they just slip. A still robot trying to push at 45 degrees would have its force reduced to 1/sqrt(2) compared to the forward direction.

Re: How Do mecanum wheels handle the bumps?
 

Aha, I was under the assumption that it's not possible to put a linear force on something via a wheel without the wheel moving forward, even by a miniscule amount ... my bad :rolleyes: . I'll ignore the fact that realistically this specific situation happens maybe one in 100 times on the field and indulge the discussion some more.

For a 100% stationary Mecanum wheel, the forces that move sideways put strain on the robot frame. Experimentally, one can see this by doing the aforementioned spinning with of a Mecanum wheel one one's hand while it touches the ground. In the experiment, it moves the arm of the person holding the wheel sideways as it rolls forward. On a 4-wheeled Mecanum robot, it simply puts strain on all of the intervening frame members creating micro stress fractures. Over time and with enough reptition, the frame will fatigue and eventually give way assuming the wheels hold up.

As such, since the sideways force is absorbed by the frame it cannot contribute to the forward component force of the robot. Since the input torque applied to the roller is split into two component forces, the forward force must be less than the input force since the sideways forces is absorbed.

Re: How Do mecanum wheels handle the bumps?
 

The frame does more more than absorb a force: It transmits the force. All these forces balance or you'll get sideways acceleration. So, a wheel has more than one input force: From the motor transmitted as a torque around the axle, and from the other wheel transmitted through the frame. Add up all these inputs and you'll get Ether's result.

Re: How Do mecanum wheels handle the bumps?
 

Yup.


~

Re: How Do mecanum wheels handle the bumps?
 

I do not insist that :-)




~

Re: How Do mecanum wheels handle the bumps?
 

This is just bad physics. The math is plain wrong as I'm interpreting it nor does it explain how a frame 'transmits' the force tangent to how the force is applied. This isn't fluid dynamics, where the pressure builds up and escapes through the path of least resistance.

If this can be proven through experimentation and then explained with clear and concise diagram, then maybe I'd consider it as valid the next time I design a robot to sit still and push a wall.

Re: How Do mecanum wheels handle the bumps?
 

Maybe I can explain it this way?

Put two CIMs into a single ToughBox.

Run the CIMs in opposite directions.

Do the forces balance? Yes.

Is there net motion? No.

Does that mean the motors won't self-immolate in a binge of stall current thirst? I think we both know the answer here. Many teams have inadvertently fashioned some very efficient OVENS out of the kit-supplied motors... it's not hard to do. Just because there is no net force, doesn't mean the energy isn't going somewhere other than motion.

Now construct a gearbox where the two motors are fighting each other at 45 degrees, but work in the same direction at 45 degrees at the same time. In one component the motors face no resistance, and all is fine and dandy. In the other component, you have the same situation as above where the motors are essentially fighting each other, and neutralizing the work done entirely to 0.

Does that shed some light into where that remaining ~30% goes?

My team is running mecanums this year, and anecdotally that 30% is noticeable - I wouldn't try and push anyone with our robot. But the benefits... they are intriguing :D.

Re: How Do mecanum wheels handle the bumps?
 

thanks a lot. i understand the concept but im having a hard time imagining how the friction plate engages with the rollers. also you mentioned that the friction plate was engaged via a pneumatic cylinder. i would imagine that pneumatics would add extra weight and complexity. wouldnt this make crab drive a better option?

Re: How Do mecanum wheels handle the bumps?
 

Crab drive is marginally less mobile, heavier, and more difficult to program and implement while simultaneously not being nearly as cool as 190's wheels.

Re: How Do mecanum wheels handle the bumps?
 

You can dissipate *energy* in a frame, but F=ma will get you in the end. Regardless of the details of your axles and struts and how the load is borne, if you have a net sideways force on the robot, the robot will accelerate sideways.

With our robot, the mechanum wheel is direct-driven by a toughbox. The motor creates torque around the axle. The rollers will convert that to a force with forward and sideways components. Our wheel is solidly attached to the axle, so the sideways force will be transmitted down the axis of the axle to the toughbox, through the toughbox mounting to a C-channel, etc. If we're not accelerating sideways, the axle is pushing back equally, with the force coming ultimately from the wheel on the other side, and by the same token, the rollers give this a forward component. For every action there is an equal and opposite reaction...

A sailboat tacking into the wind is another interesting example. The wind blows on the sail, applying a force, ignoring drag, perpendicular to the sail surface. The sailboat can then accelerate in the direction of the keel. These two projections allow the wind to blow the boat upwind.

If anyone's at the Minnesota regional later this week, we can draw some diagrams.

Re: How Do mecanum wheels handle the bumps?
 

Ether, I re-read your first post. It looks like we may be talking past each other. Seems the key is *when* the rollers start slipping, which we've yet to determine.


However, the fact remains that there are no mecanum wheels made out of traction tread. A swerve/crab drive robot with traction wheels, geared properly will beat out currently available mecanum wheels every time.

Also, mecanum wheels in motion seem to have a built in inefficiency, not all the motor's power can be applied in one direction at once.

Re: How Do mecanum wheels handle the bumps?
 

Please re-read post #20 in this thread:


~

Re: How Do mecanum wheels handle the bumps?
 

This much has been determined: In a face-to-face pushing contest between robots which are identical except that one robot has standard wheels and the other has mecanum wheels with rollers made out of the same tread material, the robot with the standard wheels will win the contest; not because of any "71% torque efficiency" issue with the mecanum wheels, but rather because the mecanum wheels will lose traction before the standard wheels do. And they will lose traction before the standard wheels do because they have higher forces acting on them (in the plane of the carpet) as explained in previous posts.


~

Re: How Do mecanum wheels handle the bumps?
 

Maybe, maybe not.

I was hoping to have a discussion of the forces/torques produced by mecanum wheels in motion after first obtaining some sort of general consensus concerning static analysis of mecanum wheels, but obtaining such consensus has proven quite a bit more problematic than anticipated.

Theoretical arguments and gedanken experiments have not been persuasive. When the season is over I hope to get my hands on some mecanum wheels and gather some additional test data. I would be delighted however if someone else beats me to it.


~

Re: How Do mecanum wheels handle the bumps?
 

Most crab/swerve systems also do not have full holonomic capabilities, and those that do require more motors and thus is a negligible weight difference. We run the in full field-oriented mode, which is much easier to implement with mecanum than with a swerve drive.

While the system does seem complex at first glance, it's really nothing more than a friction clutch system which is functionally identical to the dog clutches in most shifting transmissions used on many robots.