The past few years have set a trend for lifting games, where one can leverage the use of an elevator to not only complete basic game tasks but also an elevator based lift to complete end game tasks. Since the release of the Rev Neos, I’ve wanted to design with them, so I decided to take a shot at a Neo based elevator drive system.
During my research the fastest elevator I came across was by Team 1501, thieir 2018 Power Up elevator powered by a Vex Pro Ball shifter with 3 mini cims. By far the most powerful elevator drivetrain was Team 1678’s 2019 Deep Space triple hab 3 capable bot, powered by 4 775 in their custom gearbox.
My goal was to match the speed and power shown by both teams by leveraging the power, weight, and size advantages of the Neos. I also took inspiration from Team 118’s spool solution from their strung continuous elevator on their bot this year.
The final product is a 4 Neo triple reduction gearbox, with the first stage as a basic reduction, and the middle stage containing the shifter, with a shifter spread of 3.64x. The last reduction is an 18th:24th belt and pulley system which serves as a more effective center to center spacer for weight and space savings, and small reduction to the proper final desired gear ratios.
The gearbox outputs to a 1/2in hex shaft which contains 4 “drums”, each used for collecting string from the up or down runs for the continuous elevator, mirrored on both the left and right side for 2 parallel runs each.
If anyone knows how to make the spiral grooves seen on 118’s bot in CAD, please let me know, I’ve still been stumped by this.
Specs:
High gear: 99.2in/sec (5.99:1)
Low gear 27.2in/sec(21.82:1)
According to JVN’s calculator, it should be able to lift 550 pounds within 30amp breakers!
11.082lbs (Estimated using Inventor and Vex Pro Product descriptions)
H: 4.53in
L: 20.00in
W: 8.34in
As this is my first CD cad design post, please let me know what you think, criticism is appreciated!
I would suggest generating a helix based on the OD of the spool, with a pitch the OD of the cord, and number of turns that fills the spool space. Then sweep cut a circle profile along that path, and that circle is also the OD of the cord. You might have to play with the helix generating circle to get the correct “depth of cut”
I really like your design, and I’m interested in its development. This gearbox would allow us to speed our elevator up without losing torque to climb
On the topic of the spirals, at my work we use HUGE CNC mills and lathes (like im talking 20x20 ft large) to cut large helix grooves into drums like that (called cams), but on this scale, i would just do that on a CNC lathe or very carefully do it manually. You want to make sure to get your curvature correct on the spirals so nothing gets stuck and also grind down all rough edges.
Our team experimented with doing threaded drums, before opting to go with smooth drums. You can make them on a CNC lathe with a threading operation, or even on a manual lathe if it has an auto feed/threading mode.
Another effective way to make one is 3d printing, although you may have to use something like a markforged for it to be strong enough. I haven’t personally tested one made from nylon or pla, but I plan to later this summer.
You could also print it in PLA then reinforce it with an aluminum tube. Even with access to a markforged I’ve been making metal reinforced parts, mainly cus it prints so slow.
First, I checked your numbers with my design spreadsheet, and I got something a bit different from what you said. First of all, you’re assuming a full 12 V and 100% efficiency. In practice you probably won’t see more than 80% efficiency, and when you’re pulling 100 A continuous you’re probably dropping to around 10 V.
For your high speed, your stall load is 550 lbs, but that means if you try to lift that weight you’ll be able to support it but not actually provide any force to lift it. And to support that load you’ll be providing a stall torque of 105 A to each of the motors. On the other hand, 99.2 in/s is your free speed, meaning the speed that the elevator will move when it’s not carrying any weight (including its own). If you’re lifting 1 full robots’ weight (154 lbs), you can move at 71.5 in/s with 31 A per motor.
For your low speed, again 27.2 in/s is the free speed when you’re not lifting any weight. Now if you’re lifting 550 lbs you can move at 19.8 in/s drawing 30 A per motor.
Second, when you’re designing a gearbox to lift 550 lbs, you need to take into account some things that you normally don’t need to care about. One thing is gear tooth loading. In most FRC applications you only need to worry about shearing gear teeth in impact situations. Once you start getting forces this high though, you can actually exceed the static load ratings of FRC standard gears. By my calculations using the Lewis formula (and this online calculator) you’ll exceed the ratings of the 60t gear when lifting 550 lbs. I didn’t check others, but you probably need to switch a number of gears to steel and possibly even double up.
Another thing you need to look at is the load ratings for the belts driving the output stage. Even though you have two of them, when you’re have 550 in-lbs going through those two belts you’re likely running up against the torque transfer limit. Also you need to check shaft twisting. If your pull-up drum is on the inside then you’ll only get ~1° of twist; not so bad. If it’s the outside drum, it’s now ~7° of twist, which is more dangerous. You should actually check the shaft’s internal stress and compare it to its yield stress to make sure it won’t permanently deform, which I’m too lazy to do right now. You also need to look at the bearings, shifting setup, etc. to make sure all of them can handle this load.
Lastly, we started this season with a spiral drum. We 3D printed the outside spiral out of PLA. The inside was a core of HDPE which we bolted both to the 3D printed part and to hubs on both ends. The string was attached to a bolt threaded into the HDPE core. I believe though that we switched out the spiral cover for a flat one before the season was over. The threads just weren’t working very well for us and they weren’t needed. We were able to keep the string from wrapping on itself by using the proper fleet angle. And if it did accidentally wrap on itself once or twice the difference in elevator height is only 2\pi d_{string} per doubled up rotation, which for us was about half an inch.
What i’ve noticed from looking at other teams that did pull this off though is that if you have the string coming out of the spool vertically rather than horizontally this actually isn’t a problem as you have no force acting sideways and making the string to skip. That’s how it was done in 118’s design, as referenced by @nepa02. For our design however we needed the string to be output horizontally to allow for a pass-through elevator, so we opted for a flat spool
We printed our drum out of PLA with groves on the outside and internal hex that we printed and then broached to size. We put it on at our second regional and we never had any problems with it.
We found that, with the threaded drums, if the rope got a little off the threads, the problem would compound as the rope would begin to stack on top of itself. With smooth drums, although this would still happen to some extent, the rope was able to slide around on the spoils, allowing tall stacks to spread out and flatten instead of being held in place by grooves. Because of this, and the fact that we used very thin rope (~1/16” diameter), the problem of changing drum diameters was largely negligible.
