18-12-2009 20:52

Akash Rastogi

Re: pic: 1557 crab module

Based on side loading, wouldn't you want the other end of the drive shaft supported as well?

18-12-2009 22:00

lbarger

Re: pic: 1557 crab module

Agreed, side loading will result in some nasty torque if you use just one support bearing.

Also, you may want to look at how much area you will need for the motor to swing when the assembly pivots. Typically you want swerve/crab to pivot a good 180 degrees to maximize mobility. We stacked our motor above the wheel with a chains (or gears) going down one side.

18-12-2009 22:47

Thermal

Re: pic: 1557 crab module

On top of that, weight is a huge issue. It was stated that they are currently 15lb a module, but may go down to 12. Even 12lb a module is a hefty price for crab. You'll put 50lb on the modules alone, and still have to build your chassis to hold it on. Then you need your electronics and by that time you'll be down to under 45lb just for your manipulator. If you need pneumatics, then you're down to less than 35. It appears your axle for turning the module is offset from your wheel, which means they'll pivot awkwardly. The steel pipe is good if you want to reenact a scene from CLUE, but I don't think it has a place on a robot. Just too much weight.

19-12-2009 14:50

gorrilla

Re: pic: 1557 crab module

sorry about the late reply, I had to change my email...

again, there is only one module.

yes i know its heavy, its just a prototype. if we actually decide to make a full crab drive for competition, the pipes will more than likely be alluminum.

And the pipe is not offset, its directly over the center of the wheel. so our center of rotation is directly over the wheel.

we decided to set them up to rotate 180. and measured everything on our old frame so they stay within the boundries and take up the least amount of space.

as for the side loading, we knew this could be a problem,
for the prototype we went without it to save time, but if we make the full system, the other end will be supported. we will probably weld the entire "box" that we have now together with some extention to support the shaft.

do you think it could work without it? or is the force so much that it is a neccessity?

19-12-2009 14:59

Akash Rastogi

Re: pic: 1557 crab module

A really easy way to save weight with the Toughboxes is to replace the box tubing with standoffs.

Once you do that, you should also have some weight to play with on your prototype to add support on the other side of the output shaft. Even in the prototype stage it is smart to add the support for the side.

19-12-2009 15:11

IndySam

Re: pic: 1557 crab module

The toughbox output shaft is 4140 steel supported by two bearings already. Side loading shouldn't be a problem.

19-12-2009 15:13

AdamHeard

Re: pic: 1557 crab module

It could be. The way the plates are bolted together would allow a large enough load to deflect the two plates in relation to each other; Maybe not at first, but that force would do a great job of loosening the bolts until that happened.

I may be wrong, may be right, just my opinion.

19-12-2009 15:29

gorrilla

Re: pic: 1557 crab module

well,
we discussed how much force would be on it before we built it, and we decided it was'nt needed, but if we find it to be a problem.

the way i was thinking about supporting the shaft is, get the optional long shaft(Hex because thats what we bought hubs for) and a hex bearing, then just bolt a piece of aluminum plate to our top plate, put the hex bearing in it, and then put the shaft through it with the screw and washer on the outside.

19-12-2009 16:25

NickE

Re: pic: 1557 crab module

Sounds like it would work. You might consider removing one of the toughbox bearings if you do this. I've always been taught not to put more than 2 bearings on one shaft.

19-12-2009 18:48

nitbaj

Re: pic: 1557 crab module

The pod does look heavy, but I really like the simplicity of it. It looks like it's only 4 or 5 parts with minimal machining involved, which is nice. If you wanted to keep the pre-made gearbox idea but save weight, I would recommend going with the Toughbox Nano

As far as the side loading goes, I would say for prototyping you're fine because the gearbox does dually support the shaft and the shaft isn't too long. However, I do like the idea of putting on the long shaft and moving one of the Toughbox bearings, if possible, just to be safe.

19-12-2009 19:50

Akash Rastogi

Re: pic: 1557 crab module

Threw a rough assembly together of a nano

http://s656.photobucket.com/albums/u...ent=Assem1.jpg

I wish the CIM had enough clearance on a standard AM 4" wheel so that it could possibly be flipped into the inside of the module. Are there any COTS wheels <4" OD?

19-12-2009 20:18

J93Wagner

Re: pic: 1557 crab module

As an alternative to ToughBoxes, have you considered using BaneBots Planetary Gearboxes? http://banebots.com/c/P80K-nnnn-0005 It's what Team 93 typically uses for gearboxes. Unfortunately, I don't really know how we get them, so an e-mail to schuffsean@aasd.k12.wi.us might be in order if you like what you see.

19-12-2009 20:27

AdamHeard

Re: pic: 1557 crab module

Colsons and Banebots come to mind. I'm not a fan of the banebots though.

EDIT: With a small enough colson, you could replace the nano tubing with something a bit wider, and put the wheel inside the tubing. Sort of like 1625, but different.

21-12-2009 19:07

gorrilla

Re: pic: 1557 crab module

why?

21-12-2009 19:41

Vikesrock

Re: pic: 1557 crab module

2 points define a line. In the same way, 2 bearings constrain a shaft. Adding a third point (bearing) over-constrains the line (shaft). It is inevitable that there will be some misalignment between these three bearings. If your tolerances are very good, and the misalignment is very small, the consequences are increased shaft stresses and friction. If your tolerances are not as good, the mechanism may bind and not function at all.

Another example of this problem is a wobbly pedestal style table or chair. 3 points define a plane. When you add a fourth point, such as a four legged table or chair, you are over-constraining the design. If the tolerances are good (and the surface the object is placed on is a smooth plane) the issue is unnoticeable. If the tolerances are not tight enough, or if the table is placed on a surface that is not a smooth plane, the table or chair will wobble between two different three-point planes. A rectangular table with legs on the corners is less likely to suffer from wobbling because a virtual degree of freedom is provided by the flexing of the table surface, allowing for four constraints to all be met.