pic: Dual omni-wheel



Rollers are 3/4" diameter 60A polyurethane mounted on 3/16" inner diameter bronze bearings which roll on dowel pins. Effective diameter over the rollers is 4.8 inch. Weight as shown including aluminum hub, roller bearing, and 30 tooth schedule 35 steel sprocket is 1.75 lb. We plan to build a holonomic drivetrain using four of these wheels and demonstrate it at St. Louis Regional Fall training camp in November.

Wow. That looks really good and tough. Have you weighed it?

Hi, Mr. Wallace, this is steve (team 931, design, tall). We need to get moving on the project.

I’m curious about the actual benefit of dual wheels. Also what material are your rollers and how well do they work. If anyone has ideas for good materials for rollers that can be easily shopped i would really appreciate you sending me a message. I need something with a fair amount of traction on carpet but wont cause enough friction to cause trouble with rolling.

I posted this description with the picture (repeating here for convenience):

Rollers are 3/4" diameter 60A polyurethane mounted on 3/16" inner diameter bronze bearings which roll on dowel pins. Effective diameter over the rollers is 4.8 inch. Weight as shown including aluminum hub, roller bearing, and 30 tooth schedule 40 steel sprocket is 1.75 lb. We plan to build a holonomic drivetrain using four of these wheels and demonstrate it at St. Louis Regional Fall training camp in November.

Yeah, Steve, we need to get moving. I have a few more parts made, some gears and pinions, shafts, bushings,… but the chassis rails and other structural elements are still in work. PM or email me so we can arrange time to meet.

931 used 60A polyurethane rollers to make larger diameter (9") omniwheels for our 2003 'bot. Traction was good. Dual wheels improve smoothness, limit slip, and also provide some redundancy in case a roller gets knocked out or chewed up. Single wheels would save about 0.7 lb (2.8 for the four-wheel drivetrain) but all things considered I think the benefits are worth the extra weight.

One of the nicer omni wheels I have seen. Certainly looks more durable than many of the ones I have seen. Great job.

The TechnoKats used a similar wheel on the 2005 robot. For the front wheels we used a pair of off-the-shelf AndyMark 6" omni wheels attached together with a custom hub, similar to this. The rear wheels were simple 4" rubber wheels. Nice work!

I have just the stuff you are looking for GUM RUBBER!

Nothing beats Gum Rubber, ask anyone who has ever been in a pushing match with us!

The best thing about it is that it comes in tube form so all you have to do is cut it to length!

http://www.atlanticrubber.com/WebComp/Catalog/Public/tubing2.asp?mat=7

We used 1’’ diameter wrapped Gum Rubber on our 8’’ diameter omni wheels.

It turns just as good as any other Omni wheel out there, just with way more traction and with less bounce than omni wheels in the single configuration.

By the way nice wheels, what kind of machine did you use to make the plates?

I hope you are using them with a cool drive-train!





Mostly, they give a whole lot smoother of a ride. We used “stock” omniwheels on the kit chassis for our early software development, and it was pretty bouncy. The camera kept threatening to jiggle loose from its temporary mount.

For our “production” 'bot, we used dual omniwheels set up the same way as in 931’s picture. It drove straight as smoothly as last year’s skid-steer system, but it turned like a dream as well.

We used a water jet to cut the 6061-T6 Aluminum plates. Outer plates are 1/16" and center plates are 3/16".

We will post some drivetrain pics as the project progresses.

The extra contact points in having dualed wheels also help disperse the load for the left to right skidding. We used dual wheels like that for our holonomic ( 1083 Team Emoticons, responsible for the odd " this is how we roll " team advertisement last year at nationals. ) wheels last year ( modified andy-marks ) and it worked wonderfuly. Only problem we ran into was on the sometimes when applying power to the left and right the rubber wheels would bind up against the alluminum when a side tress was put on them. This is not good for holonomic as it provides an additional friction to overcome makes for jerky movements and difficult to keep straight due to the traction issues.

We ended up milling the slots a bit larger in order to places washers in to keep the rubber wheels firm and sliding easier. I don’t see this problem occuring with these omni wheels because of the bronze bushings placed on each side seem like they will be able to keep the rubber wheels spinning freely.

Good luck I’d love to see the finished product hopefully you’ll post some pictures. I don’t have any pictures of our modified wheels or our finished product but omni directional is deffinately a lot of fun.

Thanks for reviving the thread, Stud Man Dan.

I had been meaning to post some pictures. We did complete a holonomic drive train demonstrator, and used it as part of tutorials at St. Louis FRC Regional Fall Training Camp about a month ago. Total weight as driven was about 69 lb. Controllability on linoleum or carpet is excellent. Traction is not great, but may be adequate for a flat surface, offense-oriented game, and it should improve somewhat when weight is added.

I’ll post some more pictures tonight. One shows the anti-binding washers (ersatz thrust bearings) that you described above – we added them after I posted the picture that started this thread.

i gotta pretty newb-ish question…what is the advantage of an omni wheel over a regular wheel…is it traction? :o :o

There’s two basic ways in which omni-wheels can be used.

The first involves placing 3 or 4 omni-wheels at angles to each other. With proper motor control, this allows the robot to travel in any direction, regardless of it’s orientation - this is referred to as a holonomic drivetrain.

The second method involves a regular 4- or 6-wheel-drive drivetrain, where some of the wheels are replaced with omni-wheels. The robot still maintains almost the same amount of traction and manoeuverability (as opposed to using casters) but turning is much easier since the omni-wheels can slip sideways.

Hope that answers your question.

I hope that you also meant to ask about the disadvantages…at least for holonomic systems.
Mainly, those revolve around the force components going off in strange directions, and cancelling each other out, in order to produce a desired motion. For example, a robot with omniwheels at the corners, each at 45° to the principal driving direction of the robot, can be modelled as having force components in the direction of travel, and the perpendicular direction. Due to the 45° angle, and the fact that the wheels are 90° apart, if you drive the robot forward, the motion will be provided by the forward components, and the sideways components will act against each other (left vs. right) and do no useful work.
ΣF[sub]y[/sub] = 2 F sin(45°) + 2 F sin(135°) (Force in y-dir)
ΣF[sub]x[/sub] = 2 F cos(45°) + 2 F cos(135°) (Force in x-dir)
η[sub]y[/sub] = ΣF[sub]y[/sub] / 4 F (Efficiency in y-dir)
η[sub]x[/sub] = ΣF[sub]x[/sub] / 4 F (Effiiciency in x-dir)
For the simple case of driving forward, you get an efficiency of around 70% in the driving direction, and (of course) 0% in the perpendicular direction. As the output to each wheel varies in a holonomic drive system, the components move, but in general, either the efficiency suffers, or the total power does (as in, two wheels fully forward, and two fully perpendicular—50% power, but 100% efficiency). The same goes for mecanum wheels, in a conventional configuration.

Recently posted images. Top view, bottom view, and first test run.

Next we plan to try other roller materials, aiming for better traction than we obtained using polyurethane. Maybe gum rubber, as team222badbrad recommended earlier in this thread.

I was recently in St. Louis for a conference, and got the chance to see (and drive) this omni-drive prototype. I was REALLY impressed. It handled incredibly. If the 2006 game requires omni-directional motion, this design could be an awesome competitior on the field.

The design & construction techniques were also very cool. A lot of work and attention to detail went into this machine. I hope I see it on the field sometime soon.

Kudos!
-JV

Has anyone considered a second set of omni wheels (or a ball) in the center of the robot so the weight is evenly distributed. They need not be powered. The other thing I’m curious about is the use of different surfaces. I see some mention of using different materials and I have seen some work done on roller surfaces (the grenade wheels come to mind though I don’t think that was for traction) but has anyone really experimented with different surfaces?