pic: Omni Design Trial

[Madison]

Originally Posted by squirrel

And if you are moving straight ahead, the rollers can't roll. Even if they're loose.

It works both ways.

I'll paint half of each of the rollers and make a movie....

What's the difference between a wheel rolling forward for the sake of going forward and a wheel rolling forward for the sake of strafing and how did you get your rollers to know the difference?

[MrForbes]

Originally Posted by M. Krass

What's the difference between a wheel rolling forward for the sake of going forward and a wheel rolling forward for the sake of strafing and how did you get your rollers to know the difference?

The only difference is that the rollers are rolling relative to the body of the wheel under the strafing condition, so there is "lost motion". In the straight ahead condition there is none, the drive system is quite efficient.

A wheel turning forward is going along the straightaway of the race track...a wheel strafing is going around a curve...usually we try to go as fast as we can on the straightaway, and we have to slow down for the curves. So, a mecanum drive is naturally suited to this game, while an omni drive is not, because it is wasting a lot of motor movement spinning the rollers on the wheels.

The rollers don't care at all. But you might care if you're trying to decide which drivetrain to use, and you are still considering both omni and mecanum holonomic drives.

[Alex.Norton]

The difference is in whether or not they're constrained and allowed to move forward by the motion of the bot. In a mecanum if it is moving forward the rollers won't turn because if they were the wheel would have a lateral velocity (different from force, there is a lateral force). On an omni when moving in the forward direction there is a lateral velocity so the rollers do turn.

However this means that in an ideal situation, when you have no friction in the bearings you will not see ANY difference between the two. You may see a slight difference in top speed because of the angle on the wheels but you will see absolutly no difference in power since no power is used up in heating the bearings.

In the real world you will see that in the forward direction since the mecanum rollers aren't turning there will be a slightly better efficiency and more power will get to the floor, but I would be willing to bet that you couldn't measure the difference cause when your talking about a 321 watt motor, the amount lost to the bearings is inconsiquecial.

[Fred Sayre]

Though incredibly versatile, the standard Mecanum wheel has an unfortunate side effect which reduces its effeciency considerably. Its wide range of mobility is due to the fact that the peripheral rollers translate a portion of the motor force into a force perpendicular or at an angle to that produced by the motor. This means that are large portion of the force in one direction is lost through the translation into a resulting force by the rollers.
...
Losses of effeciency when traveling in a straight line are due to energy lost in a direction normal to that of travel through the peripheral rollers which bleed off available energy as they rotate.



From http://www.araa.asn.au/acra/acra2002...eter-Tlale.pdf

[MrForbes]

Thanks for the explanation, Alex.

[George1902]

Originally Posted by Fred Sayre

From http://www.araa.asn.au/acra/acra2002...eter-Tlale.pdf

Thank you, Fred. I don't know why I didn't think of quoting that paper earlier.

For anyone who's really interested in finding out all the science behind how a mecanum works, that paper is an excellent read.

I remember when I first came across that paper. I went to all of my friends saying, "See!? This is what I've been talking about!"

[MrForbes]

Thanks for the explanation, Fred. We now know that the maximum force that can be transmitted by a mecanum wheel in the forward direction is not as high as with a regular wheel. How does this affect the power transmission? Is there more frictional loss because of the side pointing vector? If so, how could I calculate that?

Also, what about the speed? If the rollers are not rotating relative to the wheel body as in straight ahead driving, is the full speed of the motor applied by the wheels?



The paper is interesting, but doesn't offer any explanation...it seems to be mostly selling the new design.....it would be nice if it had even a little bit of actual theory, and really nice if it had some empirical data to back up the claims.

[Fred Sayre]

Originally Posted by squirrel

Thanks for the explanation, Fred. We now know that the maximum force that can be transmitted by a mecanum wheel in the forward direction is not as high as with a regular wheel. How does this affect the power transmission? Is there more frictional loss because of the side pointing vector? If so, how could I calculate that?

Also, what about the speed? If the rollers are not rotating relative to the wheel body as in straight ahead driving, is the full speed of the motor applied by the wheels?



The paper is interesting, but doesn't offer any explanation...it seems to be mostly selling the new design.....it would be nice if it had even a little bit of actual theory, and really nice if it had some empirical data to back up the claims.

The friction comes into play of course, but that is a very hard number to get at short of actual experiments, and of course varies greatly with the type of mecanums being used.

Speed however will change. It is our experience that the rollers do in fact spin while going straight. On the ground the force is encouraging them to spin, and "perfect" traction would prevent that, but this is usually not the case. I suspect that there is some movement on the carpet, but additionally as soon as the wheel rotates the rollers will spin more as they pick up off the floor.

If at all possible, it would be a great test to get the same diameter wheels as the mecanum and compare speeds on the official carpet. We did not go into that much detail with our prototype analysis, but we made a rough estimate with vector calculations to adjust the gear ratios accordingly for our ideal fps.

[MrForbes]

So, if you were to set up the wheels to be optimized for straight driving by adjusting the friction to be high enough that the rollers would not be able to roll at normal driving loads, would that make the sideways force vector on the rollers negligible under these conditions? I'm thinking that the inability to overcome static friction would keep the rollers acting like regular tread, and the wheels would indeed act almost like regular wheels under those conditions.

I'm thinking about this as it applies to this game only.

[Fred Sayre]

Originally Posted by squirrel

So, if you were to set up the wheels to be optimized for straight driving by adjusting the friction to be high enough that the rollers would not be able to roll at normal driving loads, would that make the sideways force vector on the rollers negligible under these conditions? I'm thinking that the inability to overcome static friction would keep the rollers acting like regular tread, and the wheels would indeed act almost like regular wheels under those conditions.

I'm thinking about this as it applies to this game only.

Right, at the cost of even less efficiency going sideways. You could also adjust the roller angle to be closer to an omni wheel to increase the straight driving performance and decrease the side to side performance.

In that paper they talk about adding the feature of being able to lock or vary the angle of the rollers on a mecanum wheel to increase performance. That would be an awesome, award worthy accomplishment by a FIRST team that I hope someone will at least attempt in this or future seasons.

[MrForbes]

Or you could angle omni wheels at, say, 30 degrees to the long axis of the robot instead of 45 to achieve the same effect, right?

[M4 Sherman]

I'm kinda new with plastics, but why would you use UHMW for a chassis?

It would be virtually unbreakable, but isn't it flexible? And I though UHMW was used for more high friction situations(like chain tensioners)?

[MrForbes]

A flexible chassis is a design feature of a holonomic drive robot...helps keep all the wheels on the ground, so it will be more likely to drive the direction you intend it to.

[Bob Steele]

I just wanted to point out a small error in the discussion regarding Force and Motion.

NET force will result in Acceleration ... constant velocity requires no force whatsoever.

When we accelerate robots we see the result of all the forces involved. The net force.

The other forces on the robot are varied and the one we should discuss here primarily is friction (which is a result of the weight of the robot and the coefficnet of friction between the wheels and the carpet). This frictional force is what accelerates the robot. This frictional force is a result of the wheels applying a force to the carpet and carpet pushing back.

The net force on the robot, when it is going forward is in that direction.
If the angled wheels on a set of mecanums are not turning it is simply indicating that the sideways forces are balanced. It does NOT indicate there are NO sideways forces. I would imagine that the pin loading on the bearings in this angled state is much different than when it is moving parallel to the wheel's orientation.

I you could measure the temperature of these wheels over the course of the match I would be very surprised if they do not get quite warm.

This loss of efficiency that we all discuss with holonomic or mecanum systems has to result in energy being spent somewhere else...

Anyway these are my suggestions on this topic. Specifically they would indicate why that holonomic drives and mecanum drives have to work in the same way. Force applied does not net acceleration. Net force does.

Good luck to everyone no matter what drive system they use!!!

Have a great season!!!

May the NET FORCE be with you!!!