This is something I’ve been working on for awhile now, but never really got around to finishing. I was curious if it would be possible to fit a planetary gearbox inside a 4 inch wheel, and drive it with a NEO (or other motor.) This is what I came up with. Could use some work, and I don’t expect that it’s very practical as it is complex and would be expensive. But the package is neat.
- Wheel runs on (2) 3.25" OD thin section bearings.
- Two stage planetary gear set, all 20DP
- First Stage: 14 tooth sun, 15 tooth planets, fixed 44 tooth ring
- Second Stage: Driven by first stage carrier, 28 tooth sun, 14 tooth planet, 56 tooth ring connected to wheel
- Total reduction: 8.28:1 (12.16 ft/s free speed)
- 4" OD Wheel, 2.29" Wide
- All gears except the ring gears are off the shelf
- 5" x 5" x 4" Overall dimensions
- Second stage is intentionally 2.00:1 so that if you wanted to drive a separate wheel off of this one, you could just extend the hex shaft and use a 30 tooth to 60 tooth pulley to the other wheel.
What I originally was working on that spurred this was a motor in wheel swerve. If you used a slip ring, you could put this in a swerve too. Pivoting off the center of the wheel, the NEO would trace a 7.7" diameter circle. Still pretty compact for a swerve.
Very neat design. How is the 2nd stage carrier fixed? It needs to remain stationary so the wheel turns instead of the second stage planet gears “orbiting” the sun gear. Sorry if i’m missing something obvious.
I’m pretty sure that the screws coming out of the small planetary gears are connected to the outer side of the wheel, but I’m not sure if I’m right. Awesome gearbox any way it works.
That’s awesome, do you intend on releasing cads?
This is a really neat idea for a compact design, but I’m a little worried about the bending moment since the weight of the robot is resting on the wheel. Will the inner section bearing hold up, will there be uneven wear on the sun/planets due to the wheel trying to bend up while the motor shaft is straight, shock loads (driving off hab 2), etc.
I’m pretty sure this will stay in neutral, unless I’m missing something. The last carrier must be fixed to a non-rotating thing. Right now they look like they are fixed to that outside plate that is not fixed to anything unless there is a part fixing that outside plate that I am not seeing.
Correct, the second stage planets are fixed to the outside wheel hub plate. That’s what allows the second ring which is the wheel to rotate. The 8 button head screws on the outside plate sneak between the second stage planets, through the first ring, and into the inner hub plate that the neo is attached to. I’m realizing now that I did not add the threaded holes on the inside hub, that may be where the confusion is. Those 8 screws sandwich everything together, and fix the outer plate with respect to the inner plate.
This is a neat idea. How is the ring gear purchased or made?
It would have to be custom machined most likely. Unless you could print them in a material that could hold up to the tooth loads.
Yeah those are all very valid concerns. You could run this between two frame rails similar to the kit bot chassis. That would eliminate the overhung load issues and distribute the weight better across the two bearings.
We are working on designing a hypo-cycloidal one. That is out of the planetary family too. Advantage is less parts and just one stage. Probably after the season as an off season project. It might be easier to just make it a single stage and have the 2 to 1 reduction on the imput and keep the carrier static. Or even go 1 stage and go much faster. In your above design with a 14 tooth sun and 21 tooth planets you can get still your 56 tooth ring which gives you a 4 to 1 with the ring fixed now if you mount the input sun to a 28 tooth kinda like in the toughbox mini and drive it with a 14T on the neo you get your 8 to 1 which you seem to be looking for. Just an idea
hmm, you could fit two in <10" circle (~9.75")
(but controlling it at the module rotation still is at issue, I think it can do ~500 rpm, using center-to-center wheel distance of 5.5" diameter with the above ~12 fps. It could drop a little if only on the outside (~7.5" diameter). )
It will be interesting what works for the ring gear. Printed might work, but you have to ready for the impacts the wheels might encounter. Certainly it is simpler than machining. I believe the heat sink for the NEO is sufficient, but with brushed motors heat was a big problem for 3dp wheel modules. Printed gear teeth did hold up surprisingly well, but ymmv.
Something like that which is a prototype for an 8 in wheel the blew thing on the outside is an analog for the 8 in meccanum wheel outside. It needs a 2 to 1 on the input which I was told we do with a double helical bevel gear as soon as she figures out how to do one - also probably off season and a shrunk down version is planeed
the bolts that hold the planets are either 8mm or 5/16 and the planets have 8mm id skateboard kind bearings. the gears are as wide as the wheel. This particular one is printed in “crappy fishing line Nylon” and has an attempted 8mm bushing on the bearings printed into the wheels which does not last that well hence the next one will have reg. 8mm bearings either on the outside holding the bolts or inside in the planets.
I could see that working really well. And instead of driving a second wheel with a chain, you could have a second one of these on that side of the robot. Same amount of power as a 2 NEO gearbox, but the NEOs are at the wheel. 3rd wheel on that side could be an omni for a nice skid-steer drivetrain (If you have the frame rail on the outside, I don’t think it technically qualifies as WCD, right?).
I think this looks awesome! Dont forget to add the tread rivets or screws in the model so that your sure the gears wont clip them.
We printed a NEO wheel gearboxes based on our 2019 motor-in-wheel design using idlers. Thought I’d share the mockup (not running) so far. I’d be pretty close to 4 in width with ~4.25" treaded wheel.
We printed the idlers, but may end up with aluminium gears (which vex has in 26t).
The parts it uses (with bought idlers) are:
- 26 t 20dp spur gear (vex) 2x ~$10 each
- 10 t pinion (vex) ~ $10
- 65x85x10 mm bearing (6813-zz) ~$7 ebay
- 8x19x6 mm bearing (am-0931) ~$5
- 150 g filament ~$3.5
- Current tire is 4" intercooler coupler, but have 1" tread options too (~$5)
So around $50 without NEO/Spark. About 17 hr to print currently.
Don’t know what we’ll do with it but interesting how small it is.
Very cool, thanks for sharing! I’m assuming it’s a single stage? What is the ratio currently? I realized later with my current design that if I used 8 or 9 tooth pinions I could probably get some good ratios with just a single stage.
Right now we are going to try the 10t pinion, so the final ratio is 62:10 single stage. That give a top speed of 17.6 fps. If we switched to the 9t it would be 15.8 fps with a NEO. And you could do the 8t with the F500 at 15.2 fps.
I’ll try to add some more as the parts come off the printer of the internal build.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.