Best Form Of Six Wheel Drive?

[MrForbes]

Originally Posted by Otaku

The optimal omniwheel/traction wheel combination is omnis at the corners, tractions in the center. Rocker chassis or not, it offers the turning radius of omnis, with the traction of, well, tractions.

It depends on what you want the robot to do. In typical game competitions, the robots play offense and defense, and a robot that resists being turned (and also has to work a bit to be able to turn itself) seems to do pretty well at being able to score while being "pestered". Omnis make your robot easier to be turned around by other robots. There's always a trade-off!

I say if you're going to try to make an interesting drive system, go with mecanum. I think that a bot that moves in all directions is pretty cool.

I agree...my sons made some mecanum wheels for VEX, it's fun to drive that one around!

[crazykid234]

On 234 we went with the swampthing-esk drive where you have all traction wheels, still lowered the center wheel 1/8" with the octagon style outside. http://www.chiefdelphi.com/forums/sh...&highlight=234 We really like the "deflector shield" effect. A lot of robots like to hit other robots on the corners, but when you hit the cut-off corners the hits are kind of deflected to sides more than spinning the robot.

Downside: they are more complex to make and they weigh more.

That's just what I've seen. Like every other design there are pros and cons.

[65_Xero_Huskie]

Our drivetrain has six wheels and the key point we have in it is that our center wheels are 4mm lower than the outer wheels. This allows us to turn cleanly and to keep our robot low to the ground.

[kramarczyk]

Originally Posted by Otaku

Hmm. I've not studied these sort of things (I want to though), but that seems to make sense, except I would imagine that traction is much more than the weight pushing down on top of the wheel, although that would have a noticeable effect. You also have to keep in mind the material used as tread (Say, a plastic/perspex "tread" would be much less efficient than normal traction tread) and the driving surface matters too (such as carpet, hardwood floor, concrete, cobblestone, sand, etc.).

How then, would you calculate the traction of a robot whose weight is mostly on their right side, and the rest is evenly distributed front to back? It's a 4-wheel bot, though. (I'm asking because 675 is using 4 CIM's/4 56mm Gearboxes @ 12:1 and we can push another bot AND the rack with normal traction tread on carpet, without losing traction)

You are on the right track

How to calculate traction... well, traction is friction

(Ff) and Ff = mu*Fn

; where mu is the coefficient of friction (CoF) and Fn is the normal force acting against the friction surface.

You are correct in stating that the traction is dependant on the surfaces involved and that is what the CoF charaterizes. This information must be arrived at through testing, fortunately some wheel vendors have done the testing for us and provide those numbers along with the rest of thier product data (Thanks guys!) Keep in mind that a material may be directional and have more friction in one direction than another. Think omni-wheels.

The normal force is something that can be calculated, and for a simple case of a robot sitting on level ground the Fn is equal to the weight of the bot. If the bot attempts to climb a ramp (not sure why they would want to do that ) then the

Fn = Weight * cos (Ramp Angle)

(more detail and pics at http://www.ac.wwu.edu/~vawter/Physic...PlaneMain.html )

OK, so now what about if you have different styles of wheels on the bot?
In this case address the situation on a wheel by wheel basis. First figure out how much of the bot's weight is being carried by each wheel (possibly CAD, or several matched scales) and the appropriate Fn for the wheel. Multiply the wheels Fn by the wheels CoF to get the contribution for that wheel. After this is done for all of the wheels then add up thier contributions to determine the traction limit for the bot.

So has anyone noticed that area didn't come up? Friction, and therefore traction (for our purposes) is independant of area. For those that are getting ready to bring up dragster tires and such I suggest that you read this first. http://www.newton.dep.anl.gov/newton...ysics/PHY2.HTM
cliff notes version... dragsters use wide tires to limit wear, deal with thermal expansion, and take advantage of the adhesive like properties of the rubber compounds.

wheew...

[kramarczyk]

Originally Posted by M. Krass

This makes sense, I think. I had neglected to consider that the weight would be distributed unevenly on 'typical' 6WD arrangements since they don't use all six wheels at a time and I presume that's where the 62.5% figure comes from.

This is from the Machinery Handbook 26th ed. Keep in mind that the total load is 2W (each span has a load of W) so the reactions at the bottom need to be halved relative to the total bot weight.

I believe this attachment falls within fair use for educational purposes.

[Starke]

Working with both The Visoneers (73) and G.R.R. (340) this year in the same house allowed me to see two styles of six wheel drive. The Visoneers went with traction wheels in the middle with omnis on all fours corners. They had awesome maneuverability. G.R.R. went with six traction wheels with a lowered centered wheel. That gave them awesome traction. To answer the question which style is better depends on what you want to achieve.

[bmusser]

Our team also uses 6 wheel drive with all traction wheels. The center wheel is dropped slightly to increase turning ability. If you do not drop the center wheel you will need to run the robot with a lower gear ratio to enable turning. Unless of course you run with a ton of motors in your drive train. The other option that a few teams use is to use pneumatic cylinders with delrin blocks or small rollers on the end at each corner of the robot. When they turn, certain cylinders fire allowing the robot to skid during a turn. When the turn is done, the cylinder retracts and you are at full pushing power again. I think Wildstang has done this in the past. Good luck with the drive design.

[1359th Scalawag]

Our team uses the drivetrain with the lowered center wheel.
(hands down most popular it seems)

Last year we used just 4 wheels, one at each corner like a car. It was a terrible drivetrain. We were the slowest robot at the PNW Regional. By the end of qualifications, we were first place. Some of the judges even came up to us and said that our robot should have been losing, it was that bad of a drive train. The only reason we never lost in the qualification matches was because our robot could score all 10 points in autonomous and get onto the ramp in the end.

In the finals, the bad drive train finally got to us and we lost right away. Luckily we have a better drive train this year.

[teenmisfit]

Me and Team 501's drivetrain

[Otaku]

Originally Posted by squirrel

It depends on what you want the robot to do. In typical game competitions, the robots play offense and defense, and a robot that resists being turned (and also has to work a bit to be able to turn itself) seems to do pretty well at being able to score while being "pestered". Omnis make your robot easier to be turned around by other robots. There's always a trade-off!

Well, if somebody's going to be running Omniwheels, though, then having them at the corners as opposed to the pivot point would make more sense to me, but yes, there is a trade-off, sadly.

Originally Posted by kramarczyk

...words

Thanks for that, cleared some of it up. I'll probably get around to reading that again when it's not 10:20PM and I have to pack for a Regional.