Drive systems ideas

[EricH]

Originally Posted by team3329

We were definitely looking into mechanum for the strafing ability. So the key is 4 independent motors huh? If thats the case we have a problem. We have 4 Jags and all show that they are working with their light. however, one of them is giving the signal that it is working but it doesn't turn the motor. The others turn the motor, but extremely slowly. The battery is fully charged, so that's not the problem. is there too little power going to each one? Is there something we are over looking?

I'm not sure about the Jags, so I'll leave that to the experts. But I'd check connections if you haven't already, and see if you can get some voltage readings on the inputs/outputs.

The key to mecanums is independently powering each wheel. Quick controls explanation follows; coding it is left as an exercise for the team.
--Front/back: all wheels forward/reverse
--Strafe (side/side): Front wheels forward, back wheels reverse/front backward, back forward, depending which way you're going
--Diagonal: Wheels not on diagonal powered the same way (Annoying physics here).
--Anything in between: various combinations of the above.
--Turning: Same as a 2WD, 4WD, 6WD, etc... It behaves like a 4WD when turning, though it doesn't bounce as much as a "sticky" 4WD does.

Oh, and the other thing about mecanums: you want their footprint to look like an O, but when the wheels are viewed from the top, they'll look like an X. This is normal.

Ether has posted a whitepaper on the physics behind mecanum drive at http://www.chiefdelphi.com/media/papers/2390

[team3329]

Ok. I'll take a look at that sometime in the next few days. Ether: No, I believe we are getting just over 1/2 PWM on each, including the one that isn't working right. Would we need to change the sensitivity on the analog output relative to the joystick?

[Cyberphil]

Originally Posted by EricH

Not necessarily. The key is 4 independent motors. Chain routing can run 1 motor/gearbox to 1 wheel, which is exactly what you'd be doing with a mecanum drive.

By this, I meant one independent motor/gearbox driving 2 parallel wheels. I knew what I was talking about, I just did not convey it specific enough.

[Ether]

Originally Posted by team3329

Ether: No, I believe we are getting just over 1/2 PWM on each, including the one that isn't working right. Would we need to change the sensitivity on the analog output relative to the joystick?

What programming language are you using? Can you post your code?

[EricH]

Originally Posted by Cyberphil

By this, I meant one independent motor/gearbox driving 2 parallel wheels. I knew what I was talking about, I just did not convey it specific enough.

You mean something like the Nonadrive, but with mecanums instead of omnis?

Nonadrive: Really good drive system, but you darn well better know what you're getting into about 3 years before you even think about trying it.

[JesseK]

Originally Posted by team3329

Our team used the standard 2010 KoP drive system this year since we were rookies and it worked fairly well, but we want to try something new this year. Does anyone know the advantages and disadvantages of chain vs direst drive vs gears vs belt drives? Trying to get a big picture to try various things (Provided they don't coast too much).

Keep in mind that there are KOP-based drive trains in the championship division finals every year ( I would say Einstein, but I'm not quite sure about 2006 or 2009). For these teams, the modifications to the KOP drive train are typically gearing-based (adjustments for speed, or addition of a shifting transmission), yet the basic design stays the same. The advantage these teams is only apparent when their manipulators appear to be better designed.

Simple drive train + Elegant manipulator
vs.
Elegant drive train + simple manipulator

So my recommendation is for your team to stick with what they know, and 'experiment' with various ways to drive manipulators (belts, direct drive, etc). That will teach you everything you will want to know while also maintaining a solid, reliable drive base.

[Jared Russell]

Originally Posted by JesseK

Simple drive train + Elegant manipulator
vs.
Elegant drive train + simple manipulator

[nitpick]Why can't everything be simple AND elegant?[/nitpick]

[Chris is me]

Originally Posted by JesseK

Keep in mind that there are KOP-based drive trains in the championship division finals every year ( I would say Einstein, but I'm not quite sure about 2006 or 2009). For these teams, the modifications to the KOP drive train are typically gearing-based (adjustments for speed, or addition of a shifting transmission), yet the basic design stays the same. The advantage these teams is only apparent when their manipulators appear to be better designed.

2009 had 121, which used the KOP frame.

Simple drive train + Elegant manipulator
vs.
Elegant drive train + simple manipulator

I think this is a false dichotomy. Simple and elegant seem basically disconnected. 25's drivetrain and manipulator were both complex but worked very well this year. Most of the 2009 robots with success were as simple as possible; the teams that won Lunacy did exactly what they needed to do to win, and no more than their capabilities. This produced very simple robots like 121, 971, and 2753.


I would just like to add a reminder that the drive is the most important part of the robot. A failed experiment costs you a season, basically. This is why most teams do drivetrain prototypes in the fall when they want to try something new.

[Dr Theta]

May I also recommend you look at the FIRST WPI resource center? There are workshop and conference presentations posted there by some of the some of the most experienced members in FIRST. I know for a fact that the 2008 workshops and conferences had some very useful power points on both drive train and manipulator design.

[JesseK]

Originally Posted by Chris is me

I think this is a false dichotomy. Simple and elegant seem basically disconnected.

I claim knowledge paradox on that one: if you remember way back to when you were a 1st or 2nd year team, the two are probably like black and white. Sure, some new teams have superb mentors or students who have good experience in building machines for similar scenarios. Yet given the original poster's tone and context, I don't think that's the case.

The divide is typically made with a sense of time: if you have 3 units of time and 1 design iteration takes 1 unit of time, would you rather iterate the design of the drive train or the manipulator given that you cannot do both? Since many first and second year teams still show up to competition without even the first iteration completely built, I'd say it's a valid analogy. If this question were posed earlier in the year, there would be extra time, but alas there is not.

[team3329]

Thanks JesseK, We are actaully a rookie team, until January comes and we enter our second year. The few of us on the team that worked this year are working on teaching those that are completely new. We are kind of waiting til the game is announced before we make any crucial decisions on design because we want to focus all of our design on the game. we are just wondering so that we can have some options open for our design. We do have some experience in the design are, just not necessarily in all of these systems. I have been trying to model our team off of some of the more experienced teams that have been very successful. But thank you everyoe for your input. Geuss thats whats so good about FIRST!