Skunkworks Drive Train??
 

We are working on a new drive train.
Utilizing modified TechnoKat (Team 45 ) Omni wheels
and the BaneBot transmissions from last year.

Any ideas?

thanks

Re: Skunkworks Drive Train??
 

Team 1212 used an almost identical drivetrain (but with smaller wheels) at the phoenix regional last year and were pretty successful. It definitely provides great mobility and control, but if someone wants to push you then you are going to get pushed. I don't recall any banebots problems that 1212 had (at least in the drivetrain, I think their arm was another story...).

I like the simplicity of this kind of drivetrain; it can be assembled without much trouble at all and makes for a super swift robot.

Re: Skunkworks Drive Train??
 

I saw a design just like that at the VA Robot Rumble. They were Very mobile but got pushed around a lot. They also couldn't scale a ramp with that drive train. What if you went for traditional wheels on one axis and omni's on the other? It might be the best of both worlds.

Re: Skunkworks Drive Train??
 

Although the holonomic drive train is fun to drive and fun to build, I would NOT recommend it. The skunkworks guys should know as well as anyone how physical this game can be, i saw some of the contact they picked up in Vegas.


Many students and even mentors will say "if you are really maneuverable, you won't get hit". Thats true, but if you don't have any room to maneuver in, you will still get hit, pushed, escorted across the field, or whatever you like to call it.

My senior year, we (FRC 830) had a holonomic base. It was well built, but...

We got pushed a lot. Some of you gamers might even say we got PWN3D!!!11!!

Holonomic is a great drive system, but please carefully weigh your options.

An idea, though, would be to add a bit more support underneath. There is a lot of bowing that can occur if you aren't careful.

Re: Skunkworks Drive Train??
 

This would result in a coefficient of friction that varies with direction, and while that doesn't necessarily eliminate the drive's ability to move omnidirectionally, it does make it more difficult -- both physically and computationally.

Re: Skunkworks Drive Train??
 

Color me interested. A few questions:

1) What's the diameter on the omniwheels?
2) Why roll your own omnis over getting a set from AndyMark or elsewhere? (There are valid reasons for it, perhaps even answered by the first question, but it's nice to know.)
3) How are you attaching the omniwheels to the BaneBots shaft? It almost looks like they're floating in space.

Re: Skunkworks Drive Train??
 

Here are a few answers.

Omni and standard wheels in this configuration really would not allow movement in all direction. If we use standard wheels perpendicular to the omni wheels the omnis would have to push the standard wheels sidewise to travel in that direction. Our drive train last year had omni wheels on the corners and other drive trains I have used in the past have had omnis on one end and standard on the other. These are certainly viable configurations but do not give an unlimited range of direction without turning the robot as this drive should give.
It is simply impractical to push standard wheels sideways. Even in our drive train last year we saw the "hopping" of the robot as standard wheels were pushed in a direction slightly off the rolling axis. Completely sideways would require much higher powered motors and breakers... and of course is impractical...

We are using the wheels from the TechnoKat drawings. they are 7.7" diameter approximately. We can purchase wheels from AM and have done so in the past. We would like to make our own this year. Materials for these 8 wheels come to approximately the same cost as 8 omnis from AM...
It is the process we are interested in... One can always purchase solutions and sometimes that is the best way. We want to develop our own as part of the engineering process. The off-season is a good time to do this in our opinion. We are starting with the TechnoKat design and modifying it to suit our needs and to try and improve on the design both technically and in a way that will allow it to be made cheaper and within the constraints of our resources...

The wheels will be direct driven by the banebot motors. Application of the power will probably be done through a keyway...this will require a modification of the banebot shaft.

Re: Skunkworks Drive Train??
 

Thanks for your great comments... You are correct of course... any drive system with omniwheels on all of the corners does not give a great deal of pushing power...we will lose some efficiency due to the configuration.
When the game is released we are going to weigh our options and see if the drive train is something we want to use in competition.

The game dictates the drive train. After one has been around FIRST for awhile on realizes that the game dictates everything... which is pretty much true to a real life situation..

A big mistake that some teams make is wanting to do something because it is "cool". Although there is nothing in the rules against cool... and many super cool things are done for the game, simply using a drive system that worked before or one that is neat to design and operate are not good sound engineering decisions by themselves.

Forcing the design of the robot to work in the game is much more difficult than trying to figure out the best design for the game.

Our engineering challenge in FIRST is to come up with a machine that will best aid our alliance partners in accomplishing the task at hand. This may put us in a situation of designing a robot that isn't the drive train or the end effector that we would like or even as complex as we could design.

The constraints of the game and the resources available dictate the design.

I applaud teams for taking on design challenges in the off season and NOT using them in the game. Team 488 last year had a beautiful mecanum drive train that had been designed in the off season.. it was super cool.
They didn't use it last year.... I believe the main reason was that it didn't really fit into what they needed the robot to do. They ended up with a GREAT robot and we were lucky enough to be on the field with them as a partner during seeding rounds.

This is what the design process is about... we continue to learn....
learning is good.... challenges are good.... as Woodie says...
Humans like really tough challenges.... it brings out the best in us...

Enough for now... I love FIRST for what it does for all of us...
challenging us... making us make tough decisions.... and most of all
working with the other teams off season and during the season...

thanks for your comments

p.s. We are working on a way that we could use this design so that another robot could not push us effectively... more on that later....
perhaps much later....

Re: Skunkworks Drive Train??
 

maybe i need to clarify my comment. (i forget i'm talking to detail oriented folks) The drive I suggested would no allow "drifting" as you would see in the OP's design. It would allow the 'bot to turn within it's own radius and have the power to push along one axis. Are there better ways to do this? Yes, 6-wheel rockers come to mind. Holonomic drives are a bad choice for the games that i have seen in FIRST. There is too much contact and pushing to be feasible.I'm just not a fan of being able to apply only 1/2 my power to one direction.
I have seen very well built machines, (Henrico Triad) who use a mechanum wheel based drives, have difficultly with this years game because they cannot hold their ground around the rack. Until FIRST creates a finesse game or changes the rules to reduce the contact, power will outperform agility every time.

To Bob: L33t was not lost on me, Pwn3d... hehe.

Re: Skunkworks Drive Train??
 

I like it, for a concept. I'm not digging the shape, however. It looks like the base is a little small. You might want to arrange the modules in a rectangle (diagonally near the corners), in order to be able to use more of the allowed box.

Re: Skunkworks Drive Train??
 

That was my immediate reaction too. Also, is the picture upside-down? Building a drive-train that high off the ground might make you even more prone to tipping.

To use 254/968/60/22 (the list goes on) as an example I'd have to disagree with you. These teams have used almost the exact same system for the last 7 years with little impact on their success. The enormous quantity of experience they have gained over the years has allowed them to slowly tune and perfect the system. The same goes for any team that has used the same design every year.

These teams are guaranteed a proven, reliable system every year.

Please realize that I encourage innovation. After all, 254/968/60/22 had to start somewhere! What your team designs should be entirely dependent on your own goals for the season.

Re: Skunkworks Drive Train??
 

Your thinking is sound, but it also forces you to think differently in your approach to the game. Like Bob is saying, we had to make our robot adapt to the game, in certain situations. In 2004 we hung from the floor instead of climbing the platform. We could do so--using the arm to boost ourselves up, but it would have been hokey.

This year we choose to go with small wheels, and as a result could not get on all team's ramps. Our solution was to generally just keep scoring, or to use pure momentum and ram our way up ramps.

Sometimes it's easy to adapt something that you've been using for years, and not suffer for it. Sometimes it would be better to switch to something else.

Re: Skunkworks Drive Train??
 

Our team has designed a way to do this, yet because it is not entirely tested, we will also say more on that later. :D but it makes me happy to think about it.

I like the design, Our team was think of something, pretty much an advanced version of this design (so we are a better pushing force) but the main parts of it are the same, same gearboxes, same wheels ( 8" AndyMark Coolie Dualie Wheels to be exact.) But I like it. Nice and easy to build and gives you wicked manuverability.

Ps. I'm kinda an omni nut! lol :p

Re: Skunkworks Drive Train??
 

yes, yes you are....

Re: Skunkworks Drive Train??
 

Omni wheels for President!!

Re: Skunkworks Drive Train??
 

w00000!!!

Re: Skunkworks Drive Train??
 

2 Tips from my point of view of seeing alot of holonomic drives and researching to the extent of creating my own:

1.) Make it driver-centric holonomic. Figure out your driver's controls as well. I'd recommend having a circular nob for rotation as the Z axis and a joystick for all lateral movement. Then, when the joystick moves right, no matter what direction the robot is facing, it laterally moves right from the driver's point of view. If you push the joystick away, no matter what the robot moves down the field away from you. You'll need a gyro for this. Making a true-holonomic drive train is both hard & fun to learn, but it also can provide the agility you seek over a tank drive.

2.) Your manipulator/shooting mechanism/whatever else had better work. If your arm breaks in 2007 or doesn't work well you can't exactly go play effective defense. If your gathering or shooting mechanism in 2006 broke or didn't work you were in the same boat. Climbing ramps with omni drives this year was fairly easy *if* the ramp-bot had rails to guide you up -- otherwise you'd gun it too much and fall right off the side.

Re: Skunkworks Drive Train??
 

Fully agreed on the second part (boy, am I glad that Mike Walker talked me out of pressing for kiwi this year), though I still can't wrap my head around going driver-centric. Is it possible to do while allowing for different starting orientations? How does it work off the field?

Re: Skunkworks Drive Train??
 

It works based upon the initial orientation of the gyro. If you orient the gyro 90 degrees offset from where the driver is facing, every movement you make will be 90 degrees off unless you somehow compensate.

It's the true-vs-relative crap I do all day while coding for submarines -- it's second nature to me now, and the hard part is encoding it in a language that has no built-in geometric calculator.

Without giving too much away too early, see if you can figure out how to control it. I'll give you some hints for the smoothest, most scalable (but not THE fastest) control: use unit vectors from the gyro.

Re: Skunkworks Drive Train??
 

nice

Re: Skunkworks Drive Train??
 

I think that it is really cool how you are using our drawings! I know we put the fact that we have them avalaible to the public in our chairman's entry every year, but this is actually the first time I have heard of it being used.

A lot of Andy Mark designs are based on a design that the TechnoKats have used and has been proven to work. This is one great example as the AM omni wheels for the most part are the same design. I think it is cool again that teams are taking the "old fashioned" approach to building their robot. Even though I love Andy and Mark, I praise teams that don't go out and buy the product that anyone can use. In stead they make their own modifications and they come up with something that hasn't been made ever before. Even if it uses a drawing that has been put into production, they can never make it the same as the mass produced product.
This is the whole way that many AM products are made, as I have said before. A great example of this is our robot from last year. We wanted to incorporate the two speed transmission into our robot's frame, so we designed our own shifter using the guts of a gen2 AM. We added a couple of extras to it like an encoder and wha-la the first working prototypes of the AM SuperShifter were born a year before even being put into production. That is why I feel that more teams should go out and try to make these things on their own because that will only inspire more people to make newer better products that other teams can use.

EDIT: Also, let me know what changes you make, I will be interested to know how these are improved upon!

Re: Skunkworks Drive Train??
 

This is another great idea that is being thrown around our team in the past few weeks! We have decided that if this were the route to go, the best way to control it is with an xbox 360 controller (as many of us on the team play video games ALOT. lol) this way it makes it seem like an FPS game, which gives a driving advantage to anyone who can play FPS games well as you already have practice in! (if only we could make our robot jump... :p)