At the risk of jumping on the bandwagon, here’s a sketch for a 6-NEO drivetrain I worked on this evening. Each motor drives one wheel through a 5:1 HTD belt reduction.
The advantages are:
pretty light - each “gearbox” weighs only about 0.5 lb, including the versablock and wheel shaft.
very redundant power transmission - any “gearbox” could fail without much effect
everything fits within 5" thick x 6" high wheel wells
the chains/belts between wheels inside the frame can be eliminated if desired to save weight and increase space, as long as each NEO is speed controlled to match the others on its side
lots of room to mount the Spark Max for each NEO right on or near the motor so you don’t need a sensor wire extension
well suited for weird chassis layouts like kiwi or hexagon bases
single stage belt drive should be super quiet and efficient
The 5:1 gearing gives about 19fps free speed, which is fast for a single speed (especially if the game needs 6" wheels again). I need to see the motor curves from REV to determine whether this DT is traction limited at a reasonable current draw, or if the ratio should be higher. I can’t gear it down anymore unless I switch from belts to gears, and that doesn’t fit as well unless I move the motors, or add an idler to each gearbox, which would add weight.
having the drive pulleys on the outside of the frame rail increases the load on the cantilever drive axle. I think (without checking) that it will be find though
The biggest issue I see is that if a wheel is “floating” (that is, does not make significant contact force with the carpet or other floor), the torque does not transfer to the other wheels on the same side. You’re likely to leave a lot of torque “on the shelf”. That said, depending on the NEO numbers, this may still be ahead of many 2 CIM/3 mini-CIM drive trains at the end of the day.
Edit: Oh - didn’t see the chain/belt inboard of the tube.
I’m having a similar conversation on another forum. With 6 motors, there’s enough power to spare that it might not make a difference, and you could leave the wheels disconnected from eachother (as long as each wheel is speed controlled). Like Parthiv said, the design above allows you to tie the 3 motors together if you like, and since the power is distributed, you could probably do it safely with smaller components like 25 chain or 9mm wide belt.
Interesting design. I’m curious as one of the few elite teams last year with a 775 drive, do you think there’s any point to using them in a drive anymore?
Maybe I’m missing something, but I don’t really see the difference between this and the Single Stage Clamping Gearbox from VexPro. Other than the fact that the SSCG is designed for gears instead of belts, they seem to serve the same purpose. I have used the VexPro SSCG’s and can confirm that they hold up in a drivetrain application.
The SSCG has an 84t spacing, meaning the max reduction you can get is 11:72, or ~6.5:1, compared to 5:1 with belts. And a 72t gear has an outer diameter of 3.7”, which is no more (even slightly less) dangerous to use with a 4” wheel. It was designed for a CIM, but since the NEO has the same mounting pattern it should be compatible. And it also allows for chain on the other side of the wheel to link them together.
Is there something I’m missing that makes this design different?
Same idea yep. I prefer an aluminum solution in a drive application. I went with belts because they give less of a weight penalty for having 6 gearboxes instead of 2.
The Vex gearbox sits a little higher in the drive rail to make that gear spacing fit, which reduces ground clearance a bit, but that’s definitely a route that we could take to increase the reduction ratio
We ran a Four-CIM SSCG drive on our 2016 robot and had zero issues, I would not hesitate to run the SSCG in this configuration. Wow, a really excellent idea! I love this design.
Watch out for the tooth engagement on the pinion. It may or may not be an issue but I would be wary of that. Check this out for calculating the engagement: https://sdp-si.com/estore/centerdistancedesigner
So what is the advantage of this custom setup vs just using clamping gearboxes from VEX in the same setup?
The Clamping gearboxes can be used with a 6:1 reduction with 4" wheels, which according to my napkin math would give you an ideal speed of 16ish ft/s.
Am I missing an advantage of the belt design perhaps?
I have been thinking along the same lines, while ditching the interconnecting chains. I would make all of the reduction with staged planetaries so that the motors are coaxial with the axles. I would take all of the axle bending load via hex bearings tied to the frame rail, so that the planetaries only have to support torque. If I were designing a commercial product, I would do all of the reduction via staged planetaries, like on scissors lifts and piggyback forklifts.
To minimize leaving power “on the shelf” I would try omni wheels on one end with no drop-center.
If you don’t connect the wheels via chain/belt/etc, you can run into trouble when pushing and accelerating. The robot’s weight can shift off the front wheels and onto the back wheels, sometimes to the point where the front wheels lose contact with the ground. The motors powering the front wheels are then providing less-to-no force, and the back motors need to pick up the slack. If they are tied together, all of the motors are always taking their share of the load.
Like Matthew said, the biggest advantage of using belts vs gears as in the vex clamping gearbox is that it can be made lighter because the load is spread over several teeth. That lets you get away with a plastic pulley for instance. I thought weight savings would be important when building 6 gearboxes instead of two, but the concept is perfectly compatible with gears.
If you wanted to make a single stage planetary, you can buy (brass) ring gears from McMaster Carr: McMaster-Carr which would mesh nicely with this sun gear from Andymark using 30T planets. That would give you a 7:1 ratio if the planet carrier was the output, or 6:1 if the ring gear was the output. But that would still require more machining and would be heavier than a belt reduction.
Nice leap! Thanks for sharing — this will spark some thoughts along similar lines.
Frinstance, I am pondering 3mm belts to get more ground clearance for the driven pulley.
Modular design is a good goal.
edit: added sketch to illustrate GT2 3mm belt between Neo and 4" wheel. Parts from SDP-SI: 16T pinion and 80T pulley, machined from toothed stock. [strike]128T[/strike] belt 15mm wide. Tension using two 608 (skate-board type) bearings on 5/16 shoulder screws.
further edit: 128T is not a stock SDP-SI size, but 130T is. Looks like this layout will need to be about 1.5mm taller.
