pic: Crab Drive Frame



This is a crab drive system I had worked on a while ago, and I reworked a bit of it not too long ago and decided to share it with the CD community. The assembly is not entirely done (out of my own laziness) but the jist of how everything works is there. Total weigh (excluding chains and some hardware) is 41 lb. Not too bad for a complete and sturdy base. The blue lines are where the chains go, sadly, I a) don’t have the chain skillz of 103, and b) did this in ProE where such skillz would be rendered useless anyways; so blue lines will have to suffice. The drive is the standard AM 2-speed (no modification required) and the modules are turned using 2 window motors, mounted in the dark gray frame pieces. The frame is made of 1/8" think, 6" x 2" aluminum extrusion. The modules are assembled into the frame and rotate on 1/16th inch delrin bearings that line the inside and outside of the 6X2 tubing around the modules (notice the black squares around the left most modules). The entire system is modular, with plugs that bolt onto the side plates and slid into the front and back plates (lighter gray). These are held in by drop pins (notice in the top right). Any comments or questions!? feel free to blast me!

The thing I like the most about this frame design is that the motors are in a gear box of some sort. I seemed to of not thought of that in any crab drive frame I made, but it does seem to make things a bit neater looking and closes things off quite nicely.

Nicely done good sir!

ps. how big are the wheels? (diameter and thickness)

Thanks, the entire design is supposed to be as compact and robust as possible.
The wheels are 4" in diameter and 1.5" thick.

are you considering adding cross beams for support? (I have never made a bot with that big of a outer frame material before so I don’t know if it needs it)

Is it really such a good idea to have the drive motors sticking down that low to the ground? :confused:

You might want to invert the tranny assembly and have the motors point up. If you use it for a game with any kind of ramp, the apex might nail the motors as you try to climb.

Aside from that, nice job! :slight_smile:

Evan, you are crazy. We need to have a talk.

I don’t think cross beams will be needed, with 6" wide supports on the front and back, I doubt it will bend much. I tried to do an FEA of it to test how much force it would take to bend the frame but my computer crashed.

Low motors shouldn’t be a problem, as long as the wires are routed safely, besides they are .666" (interesting number, not by design, but curious nonetheless) off the ground, the angle of the picture is a bit deceiving.

If there are ramps in the competition, the trannys can be inverted, but they stick up pretty high, thats why I have them down.

Arefin, we don’t need to talk, you know you love this. :cool:

Thanks for the questions!

Hmm is it a good idea to have 2 gearboxes driving the wheel orientation? Maybe they move at different speeds / distance moved and cause uneven tension on the chains across the two sides…

I believe that the AM 2-speeds are driving the wheels not controling the rotation of the crab pods. The rotation of the pods looks like it’s being handled by two smaller motors. You can see the pulley/sprockets poking out of the frame on the top-left and bottom-right of the frame.

I’m concerned with the chain wrap around some of the directional sprockets (esp. what appears to be the motor sprocket). They definitely have less wrap than recommended, by whatever rule of thumb you go by. I’d be worried about chain slip. Also, how do you intend to tension the chains?

He has a “tank crab” set up right now.

EDIT: How do you plan on controlling this. Are you going to use the “tank crab” where you will have two joysticks or one joystick with some really sick programming? The picture is kind of misleading me. If you decide to link one side of your robot to the other via chain than you will save weight if you don’t want to do akerman steering. Anyways let me rephrase my question, what exactly will your crab, when it is all said and done, be able to do?

So even, i like the design, i’ve seen it before
one question;
i can see the reason for having the motors like that is so the side members can be machined the same
why not have the motor sitting in the middle? and as somebody already mentioned, it might be good to have the CIMs sitting above the frame, in case there are ramps in the game.

Arefin, stop being a crab-drive-basher. :ahh:

The main reason there are 2 motors controlling it is because you cant get a chain to wrap around the transmissions… It is really tight and is worrisome to me. Control hadn’t really been thought of, although, my friend Gustavo (software guy on 108) is really good and I am sure he’d love to get it to work with a mouse or something. In my original design, the steering chains cross (like an X) which allows the crabs to be turned tangentially, as well as a standard crab drive. I guess there is no reason not to rerun the chains in that way in this design, other than there would be even less wrap around the steering motors. I guess a tensioning system could effectively add wrap and tensioning at once (once again laziness).
Tensioning in general hasn’t really been considered, but I knew it would come up when I posted this, so right before I did, I tried to add slots to some of the holes. But for anyone who has ever used ProE, hitting the re-gen button on a dependent assembly does so with caution, needless to say, the entire thing crashed and I had to close without saving. Alas, some well placed slots and some empty sprockets (like the ones on 118) should solve the problems.

Great ideas and nice CAD, every body has a hundred ways of doing it. It doesn’t mater how you do it as long as it works for you. A thought I had with our design is to have the arm rotate with the drive directions. No mater what direction you go the arm will be going that direction. Doing that you can use one joystick, left right drives the turn motors, forward back drives the CIM’s. We don’t need to rotate the direction of the robot if the arm is always going the direction you are going. That really makes it easy, well except designing an arm that can go 360.