Gearbox in Wheel Chassis

I put this together for kicks. I’m not sure how well the addendum modified 8t pinions would mesh with an internal gear, or how well a 3d printed gear would hold up in this application. I adjusted the pressure angle of the ring gear to perhaps help with this, but I’m no expert.

It has a 8:59 gear ratio for roughly 15 fps free speed. Has a 28" by 28" frame perimeter, and weighs just under 20 lbs.

Here is a link to the CAD:
https://cad.onshape.com/documents/3ad6f9a6f80848bd8e1ea33a/w/8fd0bcc31514d764030329d5/e/b91c0ef411bfd63355c9202e

20 Likes

So it’s a cool idea, but I don’t think it’s going to hold up.

Any flex in the wheels and you’ll start skipping teeth.
Everything is cantilevered pretty hard with no support on the back side.

Being 3d printed, I’d imagine you’d shear many of those teeth before you skip any.

However, you could probably use one of those spline shaft to 8mm outputs into a 32dp pinion and purchase an internal gear like this one https://www.mcmaster.com/2696n3 . Granted, it’s a bit expensive.

1 Like

A concern with this design is that a lot of the critical parts have to be strong, small, and tight. The orange standoffs can’t be loose, the screw holding the sprockets can’t get loose or you get a sprocket messing up the inside of the tube. Lots of critical parts relying on small pieces and precision that don’t lock. With this also giving most of the torque to make the robot move, I’d add a couple more supports. Two standoffs to me does not feel like enough because it can twist. I may be wrong, but material choice may contribute a factor as to how it will handle torquing and quick changes in direction. But overall, a very cool design, very compact.

You definitely could be right here. Thankfully there are options for reinforcement, like running a tube between the gearbox plates and adding carbon fiber reinforcement into the center wheels.

I was basically just hoping that 3/4" of face width would save me here. But carbon fiber could help too.

That’s a good point. As far as the standoff goes, we have a 3/4" reamer for doing motor bosses that would probably work. The bolt on the output shaft will definitely have to be red loctited, which can quickly become a pain if it has to be removed.

We’ll probably never know if these would be serious problems, but I think it would be cool to test the concept in the off season someday. By my count it only has 27 non-cots parts, so it probably wouldn’t be too huge if a resource sink.

Maybe? But If I can shear teeth off an aluminum gear, I’d be sort of nervous about it 3dp ones. Any time you take a hard hit you could risk shock loading those teeth, which could come right off

Don’t listen to the haters, Bryce - this is ■■■■■■■ awesome.

9 Likes

I don’t see why you couldn’t make the wheel custom with wire EDM. It’ll certainly be more expensive and take longer to make, but it will be much stronger than 3D printing.

Doesn’t an EDM have to be all the way through? or am I misunderstanding how they work

I’d suggest maybe splitting the wheel module into 2-3 pieces if going that route. You would have the wheel, the internal ring gear, and the hex hub. All could bolt together. It would offer a little more flexibility in testing but more weight.

Then you could just EDM or waterjet a metal internal ring gear, if needed. Or test with other 3dp materials. Using a COTs hex hub might be another possibility, I’m not really sure how well the 3dp hex holds up (we used hubs for some pulleys we printed). Our 3dp internal ring gears held up but I’m not confident they would have with greater reduction and/or less teeth engaged. They were PETG w/ 20% CF.

Yeah, that would work. It might make things a bit complicated, but if you wanted to go that route I’d image you could do it successfully.

Also, what’d you use your 3dp internal ring gears on?

MiniCIM in wheel modules: What drivetrains are teams using in 2019?

2 Likes

That’s interesting - I would not have imagined that that would work. How much load were you putting on them?

Note that the carrier gears are static held to the motor face plate, if that was a question.

There are things that didn’t work. The gear reduction was only ~60:12 on ~4 in wheels. Increasing the ring means increasing wheel size, so there isn’t much adjustment that way. We would use a smaller pinion on future revisions to get it closer to 15 ft/s. That definitely led to some issues turning using treads and so for the 4 wheel crab drive we had to change to slick wheels, so there was lots of wheel spinning. The wheel width was ~4-5 in, but the carpet contact width because of the shape was only 1-1.5 in.

Putting all that together, along with the total robot match time (it wasn’t a great robot), I don’t think we pushed the limits of these gears much. However, in practice we did break an earlier version before adding a little more thickness beyond the bolt circle. I’m not certain if we had switched to the aluminum wheel at that point (so some flexing of the 3dp wheel might have contributed). With the mCIM there was some heat deformation of the wheel and strut, but also we could have been using a different plastic and been better.

After this year’s experience with them, would you consider doing it next year?

This seems like a concept that could really go far. Beyond 3D printing the wheels, I could see ways to do things that would work great with, say, aluminum. Maybe a machined ring gear with a mounting hole pattern that fits on the WCP pneumatic wheels (replacing the retaining ring)? In any case you’ve got me thinking about things and I love it.

Looking at the wcp wheels I think you don’t get a good enough reduction with an internal gear because the bolt circle is ~3 inch on a 6 inch wheel. However, with the above gearbox, you could slip the motors to the outside. In that case you can take a larger gear (84t+) and machine it for the bolt pattern. And you’d just be modifying a COTs gear rather than custom machining as a plus.

Why edm something you can buy?

https://shop.sdp-si.com/catalog/?brand=sdp&cid=p689

The highest ratio you can get there is a 5:1, with a 20t 32dp from vex mated to a 100t ring. At nearly $100 each, too

2 Likes