pic: Mecanum Wheel Module Gearbox

Here’s a quick mockup of the custom gearboxes driving each of the four mecanum wheels on our new drive train. These plates were cut on a Haas CNC milling machine because of the precision required for bearing press fits.

It’s obviously missing its motor, which sits inside the largest hole, as well as appropriate spacers.

Haha, quick mockup. It’s very prettiful. I less than three the Haas.

That looks really nice…even before it is completely finished.

What is the final gear ratio for that setup?

Look nice. Glad you got your machine working.

The overall ratio is 13.2:1 at the wheel achieved in two stages of 12:40 and 18:72. We have some flexibility to replace the 18 and 72 tooth sprockets in the gearbox and on the wheel, respectively, to change the ratio a bit if we’re not happy with my relatively arbitrary guessing. :slight_smile:

your “quick mockup” looks better than anything we have ever made

If that’s a mockup, I want to see the real thing…

Sweet Jesus, you guys are serious about prototyping, huh?

Looks pretty spifferific…


Looks nice. I have a few questions if you’d be so kind.

What holds the plates in the right positions with respect to each other?

If I have it right, the CIM comes in the largish hole and connects to the middle plate. Is it just the screws and the intermediate shaft that keep this middle plate in place?

How do the outside plates locate with respect to each other? Is it just the motor and the two longish bolts?

As to ratio, I assume the 12:40 is done in gears on the first stage off the motor. The 18:72 is done in #25 chain? Is that right? If so, how are you tensioning the chain? It is okay to say there is no tensioning ability, you just have to know what you are doing and do it on purpose, not by accident – which is almost always a disaster. If you don’t have tensioning, how do you say you can change the ratio by changing the sprockets – this would almost certainly require you to change the center distance of the middle shaft to the wheel shaft.

Finally, a #25 pitch 72T sprocket is almost 6" in dia. I must have something screwed up in how I think this thing works because I can’t see room for this size disk. Help straighten me out.

Joe J.

P.S. I am not picking on M here. I know that she is a pretty smart cookie. I am mostly asking these types of questions in public as a How To for folks to evaluate design concepts. JJ

All good questions, Dr. Joe. Like you, I’m interested in how the side plates are going to be spaced and aligned.

Madison has already addressed many of your other questions in an earlier thread. For detail of the gear and sprocket ratios and diameters, look here.

Thanks. I see much better now. I still have issues with how the plates are alligned. M.?

Joe J.

Attached is a plan view that will, I hope, make the following explanation make a bit more sense:


The bottommost plate on the right side represents the back plate in the photos – that which encapsulates the motor casing.

The idea was that we would be able to slide the gearbox in the slots shown on the image Richard linked to adjust chain tension. The two long, 1/4-20 bolts shown in the photograph would pass also through the plates surrounding each wheel and have a spacer that spans the distance inside the plates as shown in the plan view. Tightening those bolts compresses the outer gearbox plates against the plates on each side of the wheels which are also independently spaced to resist that compression deforming those plates toward one another.

There is a provision on the outer gearbox plates for a third bolt and spacer near the existing two that acts only as a spacer and has no function in adjusting chain tension. Similarly, there are provisions for thin sheet metal retainers to be installed over the bearings on each plate, effectively capturing the shaft and bearings from sliding out.

The middle plate will similarly be held at distance by spacers around the #10-32 bolts passing between them and into the CIM motor. There is again a provision for retainers to be installed over the bearings to prevent the shorter shaft and the bearings it rides in from coming out.

With luck, we’ll have a whole module completed, wheel and all, by Tuesday and the others by next weekend. I’m interested in seeing all the ways this will tear itself apart and fixing them now so that if we go this route in January, we’ll have a fighting chance. It’s been a long, long time since I’ve designed a custom gearbox for drive applications and I’m a bit nervous. :slight_smile:

unless you want to make the programmers tweak the code alot, then I’d suggest you’d make sure you have all the cims in the same configuration or else they will deadband in different areas… causing for different motions at slower speads

They’ll have optical encoders to play with, so I’m hopeful that’ll alleviate some discrepancy between the motors. I prefer symmetry in two planes over making the lives of my programmers easy, anyway. :slight_smile:

Yeah! What she said!

Bits weigh nothing, cost nothing, and are available on zero lead-time. And you only need two tools to work on them – the one under your fingertips and the one between your ears. :wink:

And anyway, isn’t the entire point of this chassis to be able to go in any direction without regard to the orientation of the frame? This magic is going to require tons of software. The least of the problems of the coders will be the (relatively minor) directional bias of the CIM motors.

Back to my points about spacing the plates. I think I get it. The pictures are a big help, but I don’t have the CAD in my head so I could be screwed up still.

In general, I think it is pretty nice. I like that you have options if you see flexing.

I am anxious to see the movie of your robot floating around gamefield.

Thanks for sharing your experience with us, M.

Joe J.

haha… then you’d better give them their own mock-up to play with early… (thats what our team did this year… and we didn’t have to spend the bus ride to pittsburgh programming, made for better coding, less stress, life is good. Programmers are people too! :smiley: )
you’ll kill 2 birds with one stone… they’ll be able to tweak it better, and they’ll also be able to play with the camera, and get the robot moving in the right direction.

Optical encoders --> mecanum wheels scuff. Thats what they’re designed to do, so be careful how you approach that.

Good points. Programmers also need a robot to program, which they often don’t get until it’s about time to put it in the crate!

Our team made a demonstration holonomic drivetrain in the 2005 off-season, and our programmers drove that until we got the competition robot ready. Of course they still needed some time :rolleyes: to adapt to the actual robot’s quirks. And of course they could have used more time than they got.

These are the facts of life for software engineers: (1) you will be the last member of the design team to get your hands on the system, and (2) it will be your fault if it doesn’t work.

I very much agree with that. Alot of tweaking of constants and testing and stuff have to be done … And while you are doing your stuff, everyone else puts pressure on you! Well unless part of the workflow allows the software guys to start early…

Is this the same design featured in another thread? I know the colors on the cad are the same… and it looks like you even kept the suspention idea aswell! I’m interested to see how easely this works out for you.

And did I hear someone say that the CIM motors move slower or give less torque in one direction than the other? Um… a little enlightenment please?