Elevator Carriages of Increasing Compactness

[_cmb]

Over this winter break, I’ve had lots of time to do CAD, and I’ve used that time to design a lot of elevator carriages with arm joints. My team’s robot for 2485’s CADathon had one, and now my team’s robot for the recent F4 CADathon has one (or technically two?) as well. Over this time I’ve become a bit obsessed with finding the most optimized way to package an elevator carriage with an arm joint, so I just started designing more and more, and now here I am. None of these designs involve chain tensioners, a shortcoming I am aware of and would not include on a competition robot. (Though I have been thinking about a way to include them in a future version)

Onshape document with all CAD models

Here are the four major iterations so far:

1. WCP inline bearing blocks, inline pivot

This is about as basic as it gets - simple and elegant. A NEO on a MAXPlanetary gearbox powers the rotation, with a gear reduction to an intermediate shaft and a chain reduction to the arm itself, which is mounted on a ⅞” round tube. (The specific size is because ⅞” ID bushings have a 1⅛” OD, meaning they fit perfectly into COTS sprockets with MAXSpline or standard bearing bores.) This carriage measures at 9.5” tall, and 8” wide with a 2 stage MP or 8.5” wide with a 3 stage MP*. Not bad, but it could be better…

*I decided to count the “width” of the carriages as the distance between the outer edges of the side bearings, since some of these will have different structures or bearing setups.

2. WCP inline bearing blocks, offset pivot

This is a very similar design to the first, but with the arm joint offset from the carriage. This allows the whole thing to get shorter, and gives the arm a much better range of rotation. However, because of the 0.25” aluminum plates needed to mount the axle, the carriage is 0.5” wider and is slightly heavier, though pocketing the plates helps with the weight. This carriage is 7.5” tall, 8.5” wide with a 2 stage MP and 9” wide with a 3 stage MP.

3. WCP clamping bearing blocks, 90 degree gearbox

This is where things start to get weird. Inspired by an idea from @cadandcookies, I decided to try using the new MAX 90 Degree Gearbox to turn the motor and MAXPlanetary sideways, and this was the result. The bottom box tube is completely removed, with the only structure between the two uprights now being the MAXPlanetary and the intermediate hex shaft. This carriage is 7.5” tall and 6” wide, with the arm pivot mounted slightly further out than the past version for clearance with the gearbox.

4. Custom parallel plate, 90 degree gearbox

For this version, I ditched the box tube entirely for a fully custom plate setup, mostly based on the TTB elevator bearing blocks. This allows the carriage width to shrink once again, to a ridiculous 4.25” wide. However, due to the custom bearing setup it is slightly taller than the last version at 8” tall. This was also the first version to include an encoder, although one could easily have been added to both of the first two. To make room, I removed the chain on one side of the arm and shifted the gearbox slightly to the side as well. Two chains was somewhat overkill anyway, so this works nicely, and could easily be applied to the previous version as well to mount an encoder in the same fashion. This carriage would be much more difficult to manufacture than any of the others, but it’s still possible with a basic CNC router.

Conclusion

Overall, I think version 3 is probably the “sweet spot” between COTS and custom, and still has an incredible form factor. It can also easily be adapted to fit many different COTS bearing block options. Most of the credit for this has to go to the MAX 90 degree gearbox, which is quickly becoming one of my favorite new COTS parts for the 2023 season.

As for the subject of chain tensioners, the easiest option would likely be to use a slotted connection and a cam tensioner, I just haven’t bothered to add that to any of these designs at the moment. I was also suggested by @nick.kremer to use a slotted connection for the axle with a grub screw, and I’ve been trying to think of a good way to implement that but haven’t quite figured it out yet.

[jedgar6]

I think all of these designs are cool. I’d be hard pressed not to call all of them viable. However, I’d probably steer closer to #2 personally. It’s looks more serviceable to me. If adjusted a little, all the components could be removed without much fuss…relatively speaking.

*** edit ***

With a slight bit of thinking you could just remove the vertical bearings and slide the carriage out of all of them.

[EricH]

Try putting the motor on the arm itself. Fixed (large) sprocket on the carriage. Include a chain tensioner if you can fit it. Biggest disadvantage to start with is allowing room (and slack) for the motor wires to run.

And before you ask, yes, I have seen that setup have some success in FRC.

[MaxBREAD]

Could you do the idea from version one, delete the 1:1 gear and shaft then do a single 35 chain from the gearbox and use the plate design from version 4? That would be my best guess at the most optimized while still being workable. I’m not a huge fan of the offset options personally. It’s very interesting to think about.

[jedgar6]

What about something like this. It’s a very quick but I think it conveys the idea.

(option 1)

(option 2)

Fix the two gears on either side (red arrow) for both both options. Bonus points if you cut them in half if you aren’t rotating that far. Rotate the black gear and the arm moves. Not sure how all the physics works out but It should do. With gravity I don’t think backlash would matter too much. Just have to get clever with powering the black gear on option 1.

[_cmb]

I’m working on something similar right now. Unfortunately the MAXPlanetary and 90 degree stage won’t fit inside anything less than 3" box tube, so parallel plate seems to be the solution for now. Also note that chain is generally preferable to gears for arm joints because it takes shock loads much better, which arms can experience often in gameplay.

[EricH]

Who says you need 2 gears?

I Think I Can

The video doesn’t show the full mechanism, but it shows a motor (there are 2) and the chain. Might give a bit more physics explanation that way.

[jedgar6]

I understood. Same concept but I just used gears instead of chain to make it (in my mind) smaller.

[jedgar6]

Oh, I wasn’t trying to mount the gearbox on the inside. I was mounting it to the top of the 2x1. It’s fits quite surprisingly well. As far as shearing gears…yeah I was concerned but if they are steel I’d trust them.

but yeah I think what @EricH with the sprockets in a similar configuration is probably the better option.

[_cmb]

Well this is… something.

If I fix some clearance issues it’s theoretically capable of over 360 degrees of rotation (with enough motor wires). One other problem, however, is that that the only thing holding the box tubes together is the axle, which means they could twist apart from each other along that axis really easily. That could be fixed with some simple standoffs, but that takes away the 360+ rotation, which would be no fun.

I have no idea what a reasonable use case for any of this even is at this point but it sure is cool.

[rich.varone]

I think the axel alone would be fine here as long as you were doing something like a MAXSpline, you could do a spline profile on 1 wall of the tube, then use a plug to bolt it to the outside wall of the tube. (This is potentially what you did already but I can’t tell for sure with the above picture)

[coderkevin]

Agreed, this is a really interesting use case for a MAXSpline

[SJaladi]

I think one of my favorite examples of a compact elevator carriage with a wrist joint is 971 from 2019. The use of an incredibly skinny sideways elevator enabled them to build the whole carriage from a single piece of 2x1. The tradeoff here being height for width, but in this case I think it led to a much simpler overall design given the size and tight packaging of that robot.

*Image taken from their slide deck about their 2019 robot, page linked here: FRC 971 Workshops Download Form | Spartan Robotics (FRC 971)

[Skyehawk]

Just keep in mind as you shrink a carriage (especially in height) the forces developed on other stages from torsion may result in you needing heavier materials, which in the case of mass aloft may not be a good tradeoff for the compact version space savings. Alignment of the carriage in the other stages may prove to be more of an issue as well.

All in theory of course

[EricH]

@Skyehawk

which in the case of mass aloft may not be a good tradeoff for the compact version space savings.

Oof. That video I posted earlier is a prime example. We weren’t exactly concerned about compact (there’s a lot of stuff on top of that elevator), but we were concerned about CG, after the darn thing was built. CG was high and forward–probably about 3.5’ off the ground and within 1’ of the front of the robot… which was kinda tippy anyways. A wheelie bar helped somewhat.

[GeeTwo]

@_cmb

Increasing Compactness

[Skyehawk]

[ngreen]

Our 2019 arm worked like this with the big gear fixed and the small gear walking along the chain. It is interesting that you can move the motor and gear further from the pivot it is a larger reduction. In your case like if the neo was facing inside and the gear was in it’s position with a longer chain it could be if you needed a stronger arm.

[Torrance]

Throwing the 254 2019 Elevator carriage into the mix here (see full writeup and more photos)

2 775 pros for Shoulder and Wrist, tightly packaged gearbox with side-by-side powertrains.
Pneumatic to grab onto and pull down suction climber.
Pre-suction climb it also had pneumatic actuated Dog and Tooth Comb to lockout the wrist and shoulder while climbing with the stinger.

[Orion.DeYoe]

To do this you would need a way to pivot around the MAXSpline. A 1-3/8in bushing would work just fine but I figured I’d share a little trick I found that would be good for a situation like this.

Major diameter of the MAXSpline is 1.375in which is 34.93mm. A 35mm bearing fits over this quite nicely. It’s on the loose side of a slip fit, but more than acceptable for FRC use.

6807 is probably the best bearing trade number to use. There are shielded variations (6807ZZ) along with flanged if you need it.
Here is the bearing on McMaster: McMaster-Carr