Drivetrains in Ultimate Ascent

[apples000]

Our team has been curious as to what drives teams are using on their robots. Our team's performance has never actually been limited by our drive system so this year we agreed to sacrifice a complex drive system for a more refined climber/shooter. During week 1 we played around with older robots, and determined that we would benefit from an omnidirectional drive. Several years ago our team made a shifting swerve drive that had shifters/CIM's inside the rotating wheel modules. After our full season (2 regionals/Atlanta), we never shifted to the low gear to push other teams around. This year we've done testing with 6wd with vex pro ball shifters, or mecanum wheels. We've actually found that our 8" vex pro wheels, when geared properly, let us easily outmaneuver our 6wd robot. After reading online that mecanum wheels are so bad, we were reluctant to put them on, but we've only been impressed by them. Why do people say that mecanum wheels are so bad?

[protoserge]

Our official release video will be Monday, but we used the KoP chassis and gearboxes (not KoP ratio) to maximize our schedule for more important objectives.

[bardd]

Kitbot. Simple and elegant.
I have to admit, before recieving the kit, I thought the belts won't cut it and wanted to use a frame with chains like past years, but AndyMark and Gates did a wonderful job and created the best kitbot I got to work with. This is the first time I liked working with belts.

For me, recognizing AndyMark and Gates's amazing job with this kitbot was almost enough reason for me to vote for it, even before examining its advantages.

[apples000]

I agree that the kitbot frame is a great way to go. We used almost exactly what came in the KoP! Our only modifications were some cutouts/reinforcements, and we replaced the belts with chains. Also, we've found that using well designed products from andymark is really the way to go if you need all the time you can get.

[sgreco]

Originally Posted by apples000

Why do people say that mecanum wheels are so bad?

Mecanum are "bad" because they are inefficient. The theory of a mecanum wheel is that it behave in the same manner as having omni wheels at 45 degree angles (with respect to the front of the bot) on each corner of the robot.

/ \

\ /

Sort of like that. When all the wheels spin "forward" they are generating force vectors in both the x and y direction (if you're looking down at the page, y up, x left and right). So you have a force vector from each wheel in the forward direction of Ftotal * sin45, all the x forces are lost because they cancel out. Sin45 (or cos45) is about .707. So you're only going to get about 70.7% efficiency with respect to what you could get if all the wheels were straight.

Also in order to drive "sideways" mecanum wheels have to be geared pretty low in order to have enough torque to actually move sideways. This means they will end up being reasonably slow in the forward direction in an FRC application.

Because mecanum wheels have rollers they don't have a lot of friction resistance to being pushed sideways, so they're easy to be defended. They also have trouble playing effective defense because they can be pushed out of the way sideways. In head on contact, they are limited by the 70.7% relative (and theoretical) efficiency referenced above.

Mecanum wheels also all have to be powered separately. This is not a huge deal, but it means 4 gear boxes. Then if you want to fix the low gear ratio problem and you put on 4 shifters, your robot gets really heavy really quickly.

Also, Mecanum wheels are heavy, and they come no smaller than 6 inches (unless you make you own). This means they have a high moment of inertia. A high moment of inertia in your wheels will restrict your acceleration.



(Note: I don't have any physical proof for the "70.7%", so maybe somebody has run a test and they can give you better information. Also, note that the 70.7% number is not the overall efficiency, it's a theoretical number relative to what you would theoretically get from normal wheels, all pointing straight, that have the coefficient of friction, the same weight, size and same moment of inertia).

[ice.berg]

Originally Posted by sgreco

Mecanum are "bad" because they are inefficient. The theory of a mecanum wheel is that it behave in the same manner as having omni wheels at 45 degree angles (with respect to the front of the bot) on each corner of the robot.

/ \

\ /

Sort of like that. When all the wheels spin "forward" they are generating force vectors in both the x and y direction (if you're looking down at the page, y up, x left and right). So you have a force vector from each wheel in the forward direction of Ftotal * sin45, all the x forces are lost because they cancel out. Sin45 (or cos45) is about .707. So you're only going to get about 70.7% efficiency with respect to what you could get if all the wheels were straight.

Also in order to drive "sideways" mecanum wheels have to be geared pretty low in order to have enough torque to actually move sideways. This means they will end up being reasonably slow in the forward direction in an FRC application.

Because mecanum wheels have rollers they don't have a lot of friction resistance to being pushed sideways, so they're easy to be defended. They also have trouble playing effective defense because they can be pushed out of the way sideways. In head on contact, they are limited by the 70.7% relative (and theoretical) efficiency referenced above.

Mecanum wheels also all have to be powered separately. This is not a huge deal, but it means 4 gear boxes. Then if you want to fix the low gear ratio problem and you put on 4 shifters, your robot gets really heavy really quickly.

Also, Mecanum wheels are heavy, and they come no smaller than 6 inches (unless you make you own). This means they have a high moment of inertia. A high moment of inertia in your wheels will restrict your acceleration.



(Note: I don't have any physical proof for the "70.7%", so maybe somebody has run a test and they can give you better information. Also, note that the 70.7% number is not the overall efficiency, it's a theoretical number relative to what you would theoretically get from normal wheels, all pointing straight, that have the coefficient of friction, the same weight, size and same moment of inertia).

OOOFTA!! That was a long analysis but it included all the important disadvantages of mechanums.
Well our team went with mech wheels this year for the third year in a row. But every year we change something up a little bit. Check out the live stream tonight to see what we did!

[protoserge]

Mecanum wheels are pretty awesome. They have their place, and for our strategy, this was not the year for them. Logo Motion, on the other hand, was fantastic with them.

I'm curious how VexPro Mecanums are... I might get some this summer and create a 12-CIM Mecbot

[apples000]

Our team's strategy relies on being able to move omnidirectionally, and mecanum wheels weigh quite a bit less than swerve. I think that our robot goes as fast as this game allows. We've built a few scary fast robots(15+ fps) that could cross the whole field in <5 seconds, and when driving those around, we still found that the mecanum wheeled robot driven by an experienced driver could pass by them on the way to the feeder station. We're not trying to get perfection, and we don't plan on even trying to push another robot around. Ideally, we would have swerve, but we can't afford the weight/time to manufacture, so we've compromised. I think that mecanum wheels are a great idea for average teams looking to get a quick/easy omnidirectional drive.

[F22Rapture]

Originally Posted by sgreco

Mecanum are "bad" because they are inefficient.

... snip ...

To add to this:

With mecanums, your weight needs to be equally balanced across the robot. The reason for this is that if the center of gravity is shifted to any one side, the forces will no longer balance properly and you will experience "drift." Unless your robot design minimizes this as much as possible, it will need to be accounted for.

This can be rectified by using gyros, accelerometers, and encoders to adjust the output to the wheels... but at the cost of mechanical, electrical, and programming complexity and decreased robustness.

Properly implemented mecanums can be great in games where their use is justified (high agility). However, it's something that teams should decide early on and take into consideration when deciding every other aspect of the design. It may be a better use of your team's time to optimize a simpler, more standard drivetrain instead.

Just my $.02

[apples000]

Our mecanum wheels have encoders for speed PID control, and a gyro to correct for any rotational drift. Also, we have "suspension" modules on our wheels so that they have equal force pushing down. What we've seen is that if somebody is pushing us, we have enough power to slide away, and there's enough room for our experienced driver to drive around them. We've gotten pretty good at avoiding defense.

[JohnSchneider]

Originally Posted by apples000

Our mecanum wheels have encoders for speed PID control, and a gyro to correct for any rotational drift. Also, we have "suspension" modules on our wheels so that they have equal force pushing down. What we've seen is that if somebody is pushing us, we have enough power to slide away, and there's enough room for our experienced driver to drive around them. We've gotten pretty good at avoiding defense.

You have enough power to push out of pins?

[apples000]

We don't have enough power to push out of pins, but from what we've seen, for our level of competitiveness, we don't really need to. We've built shifting swerve modules (and spent way too much time designing the shifting part), and found that when we were playing, our drivers never really had to worry about being pinned down, because they could drive quickly away from pushing teams. Our goal is to be a fast robot. Our strategy is to drive over to the other side of the field, get in the loading station, fire feeder station shots until we get defended, the zip over the the pyramid and climb for 30 pts/dump. In order to get a great climber that we felt was "safe" enough, we made some parts out of steel. To line up with the pyramid, we need to have onmidirectional drive, so mecanum wheels were our choice.

[bardd]

Originally Posted by animenerdjohn

You have enough power to push out of pins?

They have about G31 power to push out of pins.

[theawesome1730]

We are using an 8 wheel drop with VEXpro Versa Wheels. We have custom built dual output AM SuperShifters Picture here We have a low gear speed of about 5 FPS and a high gear of around 12. 6 wheel drop has always served us well, but after trying out 8 we like it better. We used mechanums once and liked them for the game (2007), but not for anything else. In other games where mechanums would have been good we opted for crab and swerve.

[ice.berg]

Originally Posted by stinglikeabee

I'm curious how VexPro Mecanums are... I might get some this summer and create a 12-CIM Mecbot

We are running the Vex mecanums and have been very pleased with the results so far. The weight advantage over the AM mec. wheels is HUUUGE. Only problem is you can not replace any of the parts on the vex unlike the AM wheels. Overall though I think it is a great product!