Re: Mecanum drive on Einstein
 

I imagine we will see some teams with mecanum excel and other fall flat. Pretty much the same as every year. Mecanum won't be completely dead after one season though, especially after a game that is such a radical departure from other FIRST games.

Re: Mecanum drive on Einstein
 

Its weird cause I hear arguments every year against mecanum but I never hear arguments that completely level them. Its just they get blown off for being mecanum

Re: Mecanum drive on Einstein
 

Pretty much. I usually ignore mecanum information in general unless it comes from Ether or a team that has used it well.

Re: Mecanum drive on Einstein
 

That is becuase there is no argument to level them. The thing about mecanum drive is that it s more or less treated as a novelty drive by most other teams (In my opinion). While mecanum may be conceptually easy mechanically, to perfect it a team needs a strong programing team to predict the exact vector of the robot one the fly. At the same time they must make it intuitive to drive. Even after the robot is done and programed you need driver training to perfect to use of this new found strafing ability. This is a large comitment for a team. Many teams like the idea of mecanum and try it with out putting the proper time or training in. This may be a reason why mecanum is so controversial. Just like any drive train a team should practice on the off season first.

It certainly has its share of pros. Most of all is the extra axis of moment which if used correctly can be used to avoid being pushed.

Re: Mecanum drive on Einstein
 

In prior years (2011 through 2014) when I was a student on our team, we had a modest reputation as having some really awesome pushing power, especially in 2013 and 2014. We ran a six wheel tank drive, and I can;t recall us ever not being able to push someone out of our way. The difference is that mecanums got pushed way easier. That being said, this year, mecanums will not be pushed around as much, so that drawback isn't present. I am of the opinion that a tank drive robot, when well done, is the ideal drive for almost every situation, and the shifting cg of this years task will make mecanums quite tricky. As much as you could use mecanum, you're probably better off spending time doing a kitbot on steroids type of base, or a custom tank drive.

Re: Mecanum drive on Einstein
 

I think a lot of people react so strongly because of bad experiences with mecanum robots. Maybe they built one that turned out to be way more difficult that they thought, or didn't work, or found that they were totally shut down by defense, or found that they couldn't/weren't using it's strafing effectively, etc.
Also, even if they have never built one, they probably have been allied with one that let them down. 2014 especially had many mecanum bots that could not dish out or take defense. Whether or not it was due to inexperienced drivers/improper implementation/poor tuning is hard to tell, so generally the drivetrain concept gets blamed.
All it takes is one bad experience to sour an idea.
Obviously, the strafing ability is very valuable. That's why we saw the emergence of Butterfly drives, more teams talking about swerves, etc.
Watching 3 Day robots makes me think more and more that your robot's ability to carefully position stacking objects will be the key feature for the game. The ability to strafe gives you one means of improving your control.
I think that this year will be the defining year for mecanums. If they don't excel this year, then they never will.

Re: Mecanum drive on Einstein
 

*I promised myself I'd stay away from the mecanum debate this year, but here goes*

And herein lies the issue with people's perception of mecanum -- there have been many teams that have done mecanum poorly. I think the reasoning for this is that it "looks easy." Teams are told what is possible and try to do it without understanding the programming/controls aspect, the physical parameters, or that any holonomic drive requires driver practice to actually be good at it.

I won't be arrogant enough to claim we've had the best implementation of mecanum, but we've done it 3 years now (once as an off season prototype), and have gotten pretty decent at it. Here are the things we do to make mecanum work well:

1. Closed loop speed control for each motor.
we attach encoders to the motor shaft directly, rather than the gearbox output. We find this gives us better, cleaner data. We do rigorous tuning of the PID to make sure we get the response we want. This guarantees the wheel is going as fast or slow as we want, rather than applying a voltage to the motors and hoping.
2. Implement a gyro for field oriented drive
the gyro allows us to know our angular position and allows our drivers to do "cool" moves like drive in a straight line while spinning. It also compensates for a wheel temporarily losing contact with the ground by doing whatever it takes to keep the robot oriented the way we want it.
3. Make an effort to keep all the wheels on the ground.
usually this is accomplished by having a frame that's not so rigid, the wheels are constrained to a single plane. Sometimes we'll actually have "shocks" on the wheels to make sure they contact the playing surface.
4. Driver practice
5. More driver practice
6. Even more driver practice


All that being said, we don't just do mecanum because we like it, we make a conscious decision each year based on what the game calls for and what we believe will work.

Re: Mecanum drive on Einstein
 

I have a secret to tell you. #4, 5 and 6 are not exclusive to any one type of drive. And it is very likely the average team that builds mecanum their first year does not build 2 robots, which puts them at a disadvantage to virtually every other 2+ year team that has an old robot with a similar drive.

Re: Mecanum drive on Einstein
 

Have you ever driven a well tuned field oriented mecanum drive? It's by far the easiest thing I have ever driven. It takes almost three seconds to get used to the "new found strafing ability," after which you literally just point and the robot goes where you want it to go. It's very intuitive.

We've also added "bearing buttons" that the driver can press so the robot can orient itself to a specific heading, even while the robot is translating.

People overestimate the difficulty of mecanum since so many teams do it poorly, but I've seen just as many poorly done skid steer/swerve/fun other drives.

Good mecanum drives are rare because mecanum wheels frankly don't make sense for most games, so the higher level teams often choose to pass on mecanum for something that meets their needs better (specifically more traction). This game is not one of those games.

For a good skid steer drive you need a strong programming team. That being said a poorly programmed skid steer drive will give you better performance than a poorly programmed mecanum drive, so it is up to every team to evaluate their resources. Mecanum is not a bad choice for every team.

Stop the h8in

Re: Mecanum drive on Einstein
 

In VEX they work really well and it is just a matter of getting the code to work. And as posted just above you need to drive, drive, drive.

With this game needing agility and not trying to push another robot 50' down field it makes sense to look at them.

I'd suggest that you get with your nearest FTC/VEX team and get them to build you a robot using these: 4" Mecanum Wheels.

If you don't have one of those then get this VEXIQ kit VEXIQ starter kit with controller and one of the Gyro sensors The VEXIQ motors have speed / rotation sensors built in so you can get all the info you need for the programming. You'll need some programming environment, there is C and Python systems.

For between $60 (local FTC/VEX) and $400 (your own VexIQ) you can build a test platform and do all the testing you want. I think you'll find that it's not that hard.

And it will help to slow down the "fear mongering" that goes around. It's really depressing to hear "we don't do mechanum since we know a roboteer on a team that knows another roboteer on a different team and they talked about doing Mecanum in 2001 and couldn't get it to work." Try a little engineering, Ether has done most of the hard work for you on the analysis.

Re: Mecanum drive on Einstein
 

Quick question:

Has anyone tested mecanums running over the noodles? Because that seriousy throws a wrench in the whole four-on-the-ground thing. I'm not saying it can't be done, it just might be an issue.

I know we drove over them with last year's drive (six-wheel tank) and had no problems driving over the noodles, although not undamaged.

Re: Mecanum drive on Einstein
 

Keep in mind that the two biggest reasons to use WCD (high fault tolerance and easy maintainence) have not changed with this game.

Re: Mecanum drive on Einstein
 

We dragged out our demo bot and tested that the second day of build season. With 6" Vexpro Mecanum wheels, running over both the standard noodle and some older more rigid noodles we had in the shop was pretty easy, although it did cause a considerable bump that might disturb an insufficiently stable stack during the noodle-crossing. There was some mild cosmetic damage to the noodle surface in some cases. The test robot was a low COG robot weighing a full 120 lbs plus battery and standard bumpers (and three frisbees).

Re: Mecanum drive on Einstein
 

Having done mecanum and then octocanum every year since Lunacy (I know, I know, boo, hiss, shall not be named and all that), I have to agree. Our octocanum drive made us an absolute beast on defense in Rebound Rumble, and one of the reasons why is because that even without field-oriented drive, the robot drives exactly like a first-person shooter--we'd use fast mecanum to get where we needed to go and then drop traction wheels to manhandle other robots.

Four encoders plus a gyro and bam, you're off and running--intuitive, simple control that most of your prospective drivers have been using for years without even knowing it. Without the gyro a human driver can compensate pretty easily for various levels of wonkiness, on the fly and without really thinking about it.

You need all four wheels on the ground if you're strafing. If you're not, mecanum drives darn close to a standard tank drive, and should have no problem whatsoever clearing the scoring platform. (We added a 'helper' wheel for ground clearance issues, but that's because of how we designed our chassis.)

My feelings in summary then are this: mecanum is mechanically easy to implement, lightweight, and is incredibly simple to drive, but requires more programming finesse than a drop-center etc etc.

Swerve seems like overkill to me this game.

Re: Mecanum drive on Einstein
 

I have not driven a finely refined mecanum drive how ever I would be open to the opportunity to have it change my mind.