Re: A new drivetrain Idea?
 

That would be cool. Whith programming any holomonic or mecanum drive could drive the way i was saying and stay locked on. That's a good point. the question is wether or not the wings would improve that feature. Time will tell with prototypes :rolleyes:.

Re: A new drivetrain Idea?
 

While the wings might improve that feature, you would have to weigh that improvement against the fact that with more moving parts and openings, the mechanical solution might be more prone to failure, as opposed to a software solution.

Re: A new drivetrain Idea?
 

The reason I don't beleive the design, as you have presented it so far, would not achieve the desired result is that it doesn't work like the skates you envisioned. In the example below, if you drive each "wing" at equal value, the resultant vector would be 45º between the two.

I think a design like yours could be acheived, and do some exciting things, but the way you have it currently wouln't work sucessfully.

Re: A new drivetrain Idea?
 

If either(but not both) of the pods had the drive direction reversed, it would turn in an arc.

Re: A new drivetrain Idea?
 

No, it wouldn't. The resultant vector would be pointing 45* into a different quadrant is all. Reverse one of the arrows Sean drew and do the vector addition again and you'll see that, without a fixed point about which to rotate, this is nothing more than one half of a typical holonomic platform. If y'all were to build this and it did travel in some sort of arc, the only likely reason for that to happen is because of inequalities in the operation of each side of the drive. The efficiences losses won't be exactly equal, so there's some chance that, given equal power, the component vector of each drive side would be slightly different.

Re: A new drivetrain Idea?
 

What if you didn't use omniwheels?

Re: A new drivetrain Idea?
 

I've decided I don't buy this. If figuring this out were as simple as adding the force vectors like you and Sean do, then I'm forced to conclude that a standard 4WD omniwheel setup can't turn. If I do your style of vector math when I drive one side forward and the other side back, I get zero and the thing doesn't move. This is patently false.

"Where's the problem, then?" you ask. Well, the forces aren't being applied through the center of mass of the robot. As this is the case, you should end up with a residual torque that will turn the robot as it translates to the side. Viola! You get an arc. However, the centerpoint won't be wherever some perpindiculars from the sides intersect. It'd depend on the center of mass of the robot, and probably a host of other factors. So, even more confusingly, it would change slightly if the CoM of your robot changed from, say, emptying a hopper full of balls.

There, I went and thought about the physics of this. On a weekend no less. I hope y'all are proud of yourselves.

Re: A new drivetrain Idea?
 

That's right. The center of gravity would represent the fixed point of rotation. Holonomic platforms have two more drive sections and those cancel the torque applied to the robot's center of mass when driving in a straight line. Because this design is asymmetric with respect to that center of mass, it will apply some torque and travel in an arc. Opening or closing the wings, however, will not vary the radius of that arc any more than it moves the center of mass along a line that bisects the angle between the wings. The drivetrain ought not represent more than, say, 15% of the robot's weight as far as I'm concerned, so you wouldn't likely see a notable shift in the center of mass.

While I certainly don't want to rain on anyone's parade and I think it's valuable for students to be entertaining and developing interesting ideas, it's also as important for them to understand what's happening to make them work and to understand the similarities between their ideas and existing technologies. It's not always fun to use what's already available (and I often don't), but sometimes doing an effective cost-benefit analysis is as valuable as implementing a new spin on an old idea.

It seems to me like Cody's been trying to examine the function of technologies like holonomic, mecanum and swerve platforms and distill from those the bare necessities that make them work. That's an awesome goal and, while there certainly are some really bloated examples of these designs, there are equally as many with important, functional features that aren't apparently obvious at first examination. Starting a design with the minimum functionality desired and working through its problems is a fantastic way of seeing how and why others have added in certain features.

Re: A new drivetrain Idea?
 

I'm slightly disturbed about the amount of half-truths being purveyed in this thread. I suggest that before making assumptions about the type of motion that a given drivetrain can achieve, you should follow Kevin's lead and review the basic kinematics and vector mathematics that goes into the actual calculations. As aspiring engineers, you should be moving past the "I think this will work this way because I think it will" and moving on to "I know this will work this way, because I've done the calculations to back it up".

For those of you who have yet to learn about these types of calculations, here are some great resources:

http://en.wikipedia.org/wiki/Vector_%28spatial%29
http://en.wikipedia.org/wiki/Kinematics

Once you've gotten a handle on those, take a look at Ian's whitepaper which discusses the calculations behind omni-directional drive systems:

http://www.chiefdelphi.com/media/papers/download/1855

With a firm background in linear algebra and kinematics, you'll find that questions like one's being posed about the arc like path, become simple and intuitive. Now, by no means is anyone expected to learn this over night, but I see a lot of energy and enthusiasm in this thread, and I have a feeling that a lot of you can digest this stuff on your own.

Remember, being an engineer isn't just about messing around in CAD.

Re: A new drivetrain Idea?
 

I probably wasn't clear enough, as I was in a rush at the time I posted.

At 90 degrees it would turn in an arc, but a very tight one, centered around roughly the same point(the pods not being at the same distance from the effective corner will cause it to not be quite perfect, I think)

The wider the angle between the pods, the larger the arc would be.

One thing that may aid in controlling the arc is a 3rd/5th motor(depending on the design) in the front as the other point of contact. Controlling the speed and direction of the extra wheel should give you precise control of the length and shape of the arc.

Re: A new drivetrain Idea?
 

I've been reading this thread lately, and it's made me quite happy. I personally believe that drive trains are often overlooked when it comes to innovation and creativity, even though they have the ability to make or break a robot. I've been spending a lot of time this off season on designing different DTs (although I'm one of the end-effector/manipulator/shooter guys), so you can understand why I'm excited to see a completely unique idea come around. A new drive train like this, whether or not it has a useful application or if it's ever used, can be used to inspire fellow DT engineers/designers to be a little more creative, possibly resulting in more efficient, exciting drive trains. So basically, I'm congratulating you for thinking outside the box and bringing innovation about for the sake of innovating. I really look forward to seeing how your prototypes work. Good luck and try to get some video! Haha.