pic: What else has Billfred been doing?

[Billfred]

Man, I thought they didn't post these things until after I finished the description! (insert anime-style sweat drop here)

Yes, it is indeed a kiwi drive. I've been working on it as part of a research grant from Preston Residential College here at USC for the better part of the semester. (Preston offered grants up to $500 to research a topic with a faculty member. It was a good excuse to learn to CAD, among other things.)

The idea is to create a kiwi drive that a team with limited resources and fabrication abilities can create. The only thing that you'd have to add on top of the list of tools needed to fabricate the Kitbot is a power drill. A center punch and/or scribe wouldn't hurt either, but you can live without them. (You can get fancier with this design--if you could find an appropriate sprocket, you could substitute an AndyMark Gearbox for the third IFI transmission and possibly shave a few bucks off of the cost. But since I couldn't find the right sprocket for the job without having one custom-machined, I kept with three IFIs.)

Most of my time has been spent just learning the finer points of Pro/ENGINEER (what USC uses) and designing those corner brackets. (You can cut two from a Kitbot's top plate. You can't really use it for its intended purpose in this setup anyway, so you might as well get something from it, right?)

We're hoping to get this thing together for Discovery Day (USC's annual undergraduate research shin-dig) in a couple of weeks. It should be fun.

Originally Posted by Greg Needel

i hope you are going to have some ball casters or skids under the corners of the triangle or it might be unsteady.

Something will be going under there. I'm thinking PVC, since I already had to buy some to make spacers with.

Originally Posted by Stuart

I personaly would move the wheels to the corners instead

It could be done like that, but one of the first things I learned from my mentor (Dr. Philip Voglewede, if anyone was at the Palmetto team social) was that the wheels's axes needed to converge in the middle for things to work out smoothly. To get them in the corners would take, at least in my judgement, a lot more trouble than it's worth. (But if someone has an idea, I'm all ears.)

[Stuart]

Originally Posted by Billfred

It could be done like that, but one of the first things I learned from my mentor (Dr. Philip Voglewede, if anyone was at the Palmetto team social) was that the wheels's axes needed to converge in the middle for things to work out smoothly. To get them in the corners would take, at least in my judgement, a lot more trouble than it's worth. (But if someone has an idea, I'm all ears.)

it is an equilateral triangle isn't it? . . . cause if it is just moving the wheels to the coroners would keep the axises centered. . . then for the kit drives just move them above the wheels instead of to the side of them. . . . this is all of course assuming that it is an equilateral triangle

[Tristan Lall]

Billfred, I think that people are suggesting that the wheels be mounted to the triangular brackets, with their axes perpendicular to the long edge of the brackets. That allows the robot to be fully supported on its wheels, rather than causing some of the weight to rest on undriven wheels/casters/skids (which would reduce the amount of force that the robot could exert in a pushing match).

Also, why not consider a quadrilateral version of the Kiwi drive? Like what 854, 931 and others used this year? It seems even simpler to build, and a little more powerful (provided that the extra CIM wasn't needed elsewhere).

A note on the kitbot sprocket. I just realized that we ought to be looking for ISO 04 and ISO 06 sprockets, not #25 and #35. They're functionally equivalent to ANSI roller chain sprockets, if I'm not mistaken, but they do come in 16 mm bore sizes. In fact, SDP/SI has some, in larger tooth sizes for ISO 04/#25. And Martin sells a few ISO 06/#35 ones with 16 mm bores. Search GlobalSpec for "Sprockets, Roller Chain, Metric", for more sources.

[Billfred]

Originally Posted by Tristan Lall

Billfred, I think that people are suggesting that the wheels be mounted to the triangular brackets, with their axes perpendicular to the long edge of the brackets. That allows the robot to be fully supported on its wheels, rather than causing some of the weight to rest on undriven wheels/casters/skids (which would reduce the amount of force that the robot could exert in a pushing match).

The angle of the drawing (and my comments) might be unintentionally tricking folks--it does rest solely on the omniwheels. Any skids I might add to the corners would be just for some extra insurance--a wheelie bar of sorts. (Though don't most kiwis attempt to avoid pushing matches anyway?)

Originally Posted by Tristan Lall

Also, why not consider a quadrilateral version of the Kiwi drive? Like what 854, 931 and others used this year? It seems even simpler to build, and a little more powerful (provided that the extra CIM wasn't needed elsewhere).

Why not quadrilateral? Good question. Besides a desire to do something a little different (I only know of 857's in 2002 and 229's prototype kiwis in FIRST, but there have been many more quadrilateral ones), a three-wheeled version has an additional thing going for it: cost. You don't have to buy or make the fourth omniwheel/transmission/etc. (Besides, that also frees up a CIM motor for other things, which I'll leave to the reader's imagination.)

Simplicity shouldn't be too much of an issue--those corner brackets are the sole part you have to custom-fabricate, and that's just a matter of lining up your hacksaw correctly. Everything else is just assembling the Kitbot with the low-rider configuration (although you could go the traditional way if you desired--I just went low-rider for the center of gravity benefits.)

Originally Posted by Tristan Lall

A note on the kitbot sprocket. I just realized that we ought to be looking for ISO 04 and ISO 06 sprockets, not #25 and #35. They're functionally equivalent to ANSI roller chain sprockets, if I'm not mistaken, but they do come in 16 mm bore sizes. In fact, SDP/SI has some, in larger tooth sizes for ISO 04/#25. And Martin sells a few ISO 06/#35 ones with 16 mm bores. Search GlobalSpec for "Sprockets, Roller Chain, Metric", for more sources.

Interesting. I'll have to check it out.

Originally Posted by Tim Delles

Also don't run it on any hard floor, only use carpet. If you run it on a hard floor it makes the most horrible sound in the world. Well close to it

Something to keep in mind--luckily, the halls of both 300 Main (the mechanical engineering building) and Preston are carpeted. Fwahaahaa.

Originally Posted by Stuart

it is an equilateral triangle isn't it? . . . cause if it is just moving the wheels to the coroners would keep the axises centered. . . then for the kit drives just move them above the wheels instead of to the side of them. . . . this is all of course assuming that it is an equilateral triangle

It is an equilateral triangle--the corner brackets only took me about 47 tries to get right. The wheels are mounted where they are for simplicity and durability (how many decently-constructed kitbots have you seen smashed to pieces?). This setup is basically just the lowrider configuration with some interesting corners. I haven't fully investigated mounting them on the corners, but I can't picture it being done in a cheap-and-easy fashion.

Originally Posted by sanddrag

If those were 6" wheels (they look like 8" right now, are they?) you could get away with putting them right on the gearbox output shafts. Although you'd need a custom hub, which is not desired in this project I guess.

Those are 6" AM omniwheels. I didn't plan on traversing anything other than flat carpet, so I didn't see a benefit to spending $12 more for 8" wheels all around. You're right, though--the hub was undesired.

I appreciate the feedback--it's good to hear it from a lot of folks.

[JVN]

The 229 version that was linked in this thread was designed primarily by Chris Carnevale.

Chris' design consisted of designing a simple "Omni-Module" which was a solid unit consisting of one AM omni-wheel, 1 gearbox (he used the Bosch Drill+Gearbox, but a kitbot gearbox could easily be substituted), and all the mounting/framing that comes with it.

Chris then built 3 of these simple omni-modules, and bolted them onto a big triagular piece of plywood. Simple, and easy; this design can definitely be built by low-resource teams.

-JV