An Answer About Frc Elevators

[TL;DR at the bottom]

Hi everyone! Considering how the last time we had an elevator game was in 2019, I’m now the only member left on my team that has designed an elevator (2019 was my team’s rookie year and I’m the only member that was on the team that year that is still on the team since everyone else has graduated, and even our current mentors joined the team after the 2019 season). Fortunately, I was in charge of designing the elevator back then as well, so I figured I would take this offseason to design and build an elevator, because even if we don’t have an elevator game next year (which, honestly, kind of sick of shooter games, I want an elevator game again), I should at least teach all the current members how to build an elevator for future seasons.

Primary design constraints:

• I designed it to fit within the constraints of our 2019 robot, so 20” wide overall and 42.5” tall (to account for the 2” that would be lost under the elevator with 6” wheels).
• This is also three stages, which in my mind is the max I would ever need. It reaches well over both the scale from 2018 and the rocket from 2019 (7ft maximum on both, 9ft maximum on this).
• We have access to a 3x4ft CNC router and two Markforged printers (an Onyx One and a Mark Two), and a Tormach 1100 PCNC (not used in this project)
• For simplicity’s sake, I designed this to have as little “milling” on the router as possible, so aside from the mounting plates for the spool shaft, it’s only drilling holes so you could probably do almost all of this with a drill press.

Now, with that said, here it is!

image image

Main features:

• 3 stages
• Cascade rigging
• Two sets of 1/8” UHMWPE cable rigging
• Pulldown and pullup ropes
• Spring tensioned (not modeled)
• CF spring counterbalanced (only set for 10.6lbs each, but could swap these with heavier springs if needed)
• 1 or 2 NEO / Falcon drive at a 10:60 gearing (two motors allows for the ability to climb)
• Currently, can reach up to 108” off the ground while still having 8” of stage overlap
• 1/8” gap between stages
• Current weight: 21.31lbs, plus another ~0.2lbs in cables and ~0.3lbs in cable tensioning springs, and ~0.1 in rivets. Sub 22lbs overall.

image image image
• All of the bearing blocks for our elevator are printed out of Onyx in two halves, then put together with ½” OD #10-32 threaded hex standoffs from McMaster-Carr. This allows for the bearing blocks to be structural and double as both bearing blocks and structural components, and the hex keeps the standoffs from spinning inside the blocks.
• Uses ¼” ID, ¾” OD bearings for the 1” side of the elevator and ¼” ID, ½” OD bearings for the 2” side.

Rigging
image image
• The elevator is Cascade rigged, which I did mostly because it’s what I’m familiar with. You could switch to Continuous if you wanted to, probably (?) wouldn’t be too difficult, you would just need to move the pullies around.
• Currently, I have twice as many pullies as needed because each cable run is connected with it’s mirror on the other side by a spring in the middle, tensioning both sides at the same time and removing any slack caused by differences in the two sides. You could use just a round pin to turn them 90 degrees instead of these pullies, and I may switch to that later for cost reasons.
• There is also a pair of constant force springs being used to counterbalance, which are mounted to the base stage (as in, the unmoving stage) and slips between the base stage and the first stage and connects at the bottom of the first stage.
• The moving part of the elevator only weighs 11.5lbs currently, so assuming a 10-15lb intake, the two 10.8lb springs are all that should be needed to counterbalance the elevator.
• Really, even with one NEO, the CF springs aren’t needed and I most likely wouldn’t end up using them, just so I don’t have to deal with the potential safety concerns.

Spool
image image image image
• The spools are, at least in this configuration, driven by two NEOs on a 10:60 reduction, which allows the elevator to reach full extension in under a second while still being able to climb.
• You could easily switch to a 11:60 or 12:60 (since the pitch diameter of those pinions is the same), use Falcon 500s, or switch to 1 motor if you weren’t climbing.
• The spools are ¾” in diameter (for the center of the rope, so effective diameter in terms of torque) for the pullup ropes and 2 ¼” in diameter for the pulldown.
• The difference in diameter is because I am pulling down from the carriage, instead of the first stage, so the pulldown rope moves 3x as fast as the pullup rope. If I were to do a 2-stage elevator (which is probably more likely given that neither 2018 or 2019 really needed a three-stage elevator), I would probably do 1” and 2” diameter drums instead.
• The spool shaft is held by a frame made from 1/8” plate and 1x1x0.040” versaframe connected with 1/16” gussets. Most likely, on a real robot, I would just do the brackets on each end directly on the drivetrain rails for simplicity.
• The two NEOs and the gears are in the center of the shaft to reduce twisting or flexing along the length of the shaft

Cost
image
Note: This order form was put together for my team, so it excludes things like fasteners and motors we already have on hand and includes spares for all parts. This is an order form, not a bill of materials, but if someone would like, I could put together a more complete bill of materials.
• The overall cost of the elevator would come out to about $1000 for us, including shipping.
• The main costs are Onyx for the Markforged printers and Versaframe tubing from VEX, with the rest of the money going towards the various other COTS parts involved.
• You can just do normal 1/16” walled tubing which would save about $130 at the downside of weighing about 3.5lbs more in this elevator configuration. Considering the high center of gravity of most of this weight, especially when the elevator is extended, I considered this cost to be worthwhile.
• PLA or PETG would likely work, just would need to take some time to adjust all the fits due to print shrinkage, and you may need to beef up the parts a bit more to account for the potentially reduced strength. (we did use the inlaid continuous fiber from the Mark Two on two parts, the brackets for the rope pullies and the constant force spring, so I would be more cautious with printing those out of other materials. Fortunately, you could probably slightly redesign those to be made from normal angle iron.)
• You could also reduce the number of pullies used by about half, saving ~$40, if you used a round dowel instead of a pulley to redirect the rope for tensioning.

Thank you! Let me know if you have any questions or comments, or if there’s any other ways this could be modified or improved. Best of luck to everyone!

This one’s for you, Rohit:
image

[DR;TL at the top]

LINKS
Order form: Order Form 2021 Offseason - Google Sheets
Solidworks/STL files Download: Custom Elevator - Spring 2021 - Google Drive
STEP File Download: https://drive.google.com/file/d/1G7L2N8XRK6MHHMW6LW9qZp96Hpce3HM9/view?usp=sharing

48 Likes

11/10 title

7 Likes

Completely independent of the quality of the elevator itself, I want to commend you for such a comprehensive post.

You presented all the necessary steps to designing and justifying your design decisions covering the following:

  • Requirements / Design Inputs
    Overall dimensions
    Stroke / reach / range of motion
    Manufacturing capability

  • Key design decisions
    Cable tensioning
    Gaps / clearance / tolerance considerations
    Number of motors, their reduction, and resulting speeds/torque
    Why using 3D printed mounts and justifying why you believe they’ll be strong enough

  • Output / final design summary
    Screenshots of final CAD
    BOM with cost breakdown
    Design-for-manufacturability analysis on the trickiest parts
    Mounting to frame / system-level integration
    (Next step would be part drawings if design is approved)

I do think with a few more section-breaks or list-formatting you could make some of these paragraphs easier to read.

Overall, however, this is a great lesson in many of things I’d expect a fellow engineer at work to present, quite impressive coming from a student. Would definetly earn high marks if I was judging this in a CADathon or interview panel like Chezy CAD.

27 Likes

Neat stuff, very thorough.

Would you mind drawing out how the rigging would look on the elevator if possible? Always have trouble understanding that part since no one really models it.

Also, have you built this (the wording makes it seem like you did)? If so, can we see some pictures?

1 Like

Thank you for the suggestions and feedback! I made this with the hope that someone else would be able to get some helpful info out of it, whether or not anyone builds it, so I’m glad to see I got all the relevant information across.

We have not built this elevator, the hope is to build it later this offseason, although due to budget reasons, I would expect we probably do not. I would like to make some of the elevator blocks at the very least just for proof of concept, but we will see. I will keep y’all posted.

As for explaining the rigging, I’ll try my best…
Here’s a screenshot of the CAD with some lines drawn. There is a spring in each set of rigging, indicated by a black box. Also, goes without saying, some artistic liberties were taken in order for the picture to be more readable. In the real CAD, almost all of the cables are straight up and down, because having angles can lead to inconsistent travel.

Imgur

Red lines: pullup cables. Connects to the spool and goes up to the pulley at the top of the base stage, and then connects to the bars at the bottom, pulling them straight up.

Green lines: second stage cables. Connects to the bottom of the second stage then to the pulley on the top of the second stage, then connects the other end on the base stage with a pulley to turn the rope 90 degrees for the tensioner. As the first stage is lifted, it pulls the second stage up because the pulley on the second stage pushes up on both the end going to the bottom of the second and the end connecting to the base stage - the base stage can’t move, so it pulls on the end going towards the second stage and therefore lifts it.

Yellow lines: third and final stage cables. Same as the Second stage cables, except with different stages. it connects to the third stage (the carriage), then goes up to a pulley on the second stage, before ending with the spring tensioner on the first stage. Second stage rises, pulls on both the first and third stage, first stage can’t move because it’s what’s pulling up the second stage, therefore it can only pull up the third stage.

Blue lines: pulldown cables Connects the larger pulldown spool directly to the third stage, the carriage. This feeds out line 3x as fast as the pullup spool as the elevator is rising, then pulls the carriage down when the direction of rotation is reversed.

I hope this helps! One of the disadvantages of cascade is that it’s harder to understand, meaning it’s easier to make mistakes. Messed up the rigging several times back in 2019. May actually consider buying several different colors of cable just to make it easier to follow.

14 Likes

Awesome.
Congratulations, and I look forward to seeing one in action!!!
Thank you for sharing - I know this will be an inspiration for all design teams.

2 Likes

Agree with excellent quality of post. Does your school offer technical writing as an English course :smiley:

1 Like

Hi everyone! Small update with the elevator, yesterday I came up with an idea that is, admittedly, very overkill and solves a minor problem, but I’m still happy with the result and thought it might be worth sharing.

In short, in a normal spooled elevator, as the drum spools or unspools, the angle at which the rope goes from the drum to the first pulley changes. This is possibly problematic for a few reasons:

• The angle changes, which slightly changes the amount of lift the elevator gets for each rotation
• Can result in the coils starting to overlap and pile up, which effectively changes the diameter of the spool
• In extreme cases, it can cause the rope to jump the pulley or the spool completely

Both of these result in a different amount of cable per turn being released, which by itself isn’t an issue, but this does result in a possible discrepancy between the four spools (two lifting, two pulling down), which can bind the system and cause other problems. Now, I do have a spring in each set of rigging which would most likely negate this issue, and placing your spools farther away from your elevator helps as well, but what if that’s too simple for me? Therefore, I came up with a mechanical solution to ensure the cables always feed at the exact same angle.

Behold! Leadscrew!
This spool has a leadscrew on one end with the exact same pitch as the spools, so at the same rate as the spools unwind, this moves the whole shaft to the side, therefore keeping the cables at the exact same effective angle, both when pulling the elevator up and down! This works by having a coaxial leadscrew that feeds into a stationary nut, therefore moving the leadscrew back and forth.

Pros:
• Peace of mind?
• Allows you to place the spool as close as you want to the elevator, opening up potential options for packaging
• Fits in mostly the same footprint as without the spool (same width and height, just longer), or can even possibly be added on to an existing spool if needed
• Cheap; under $50 extra, VEX shaft coupler costs more than the leadscrew and nut combined
• Relatively simple - does not need any complex manufacturing, adds about half a dozen major components. Leadscrew nut block could be 3D printed, no problem.

Cons:
• Slightly more complex compared to not having it, which leads to this being slightly harder to manufacture and assemble, and more possible points of failure
• Longer - this takes up 28.5" of width just to make sure the spools are centered in the robot and in a place where they work well with the elevator.
• If the elevator was much wider than 20", you’d have to essentially move the pullies in relative to the elevator itself. This is likely fine normally but means the rigging could interfere if you were doing a passthrough on your elevator. The original spool without the leadscrew comfortably fits within the profile of the elevator with no problems, regardless of width.
• Slightly heavier; the whole spool weighs 5.80lbs now, compared to 5.15lbs before, so that’s a 0.65lb addition - not insignificant. This alone would make this not worthwhile to some teams.
• Works by sliding the shaft back and forth in the thunderhex bearings and the center gear - not desirable and not an intended use case for the shaft or bearings.

Overall? I’d say, could make sense for some teams, especially if packaging forces you to place the spools very close to the elevator. Figured there was a small chance a fellow FRC team could get some use out of this. If we do build this elevator in the offseason I’ll give this spool design a shot and see if it works at all. Thank you!

10 Likes

I will point out that in 2018, it was important to build an elevator that could not only place cubes when the scale was at 7ft but place cubes on cubes when the scale was at 7ft. Take it from a team that didn’t do that. Your elevator looks great and definitely exceeds that.

Oooh, I love the differential windlass - saves the weight of a gearbox (if you can make room for the drums)

Edit - wait, am I wrong? Is the other drum for retraction and not a differential spool?

1 Like

Sliding a shaft through a bearing bore while it is loaded is definitely not a good idea. Without having tried it I would say that this is a recipe for binding. Plus, you have substantially reduced the amount of support that the shaft is getting. The shaft coupler will not be as rigid as a solid shaft and the outermost supports are now further away leading to more shaft bending under load.

As an alternative, you could design a guide or fairlead system for the rope itself that shuttles back and forth using the same leadscrew concept. The leadscrew could be on a parallel shaft to the spool shaft and would move the guide back and forth.

In general, we have not found a need for this. As long as when the spool is wound so that the end of the pulldown rope that is attached to the third stage is in line with where the rope is winding on the pulley when the stage is all the way down, the angles that you get when the third stage is all the way up are very small and there is little risk of overlaps or skipping grooves on the drum.

We have found, however, that there is a need for a rope tensioner in the system. In your design, you could attach your pulldown rope to the upper crossbar of the third stage which would give you the space between the upper crossbar and the spool when the elevator is fully down to insert a tensioning spring. I have also seen an arrangement where the pulldown rope goes through a pulley on the third stage to turn the rope sideways and then install the tensioning spring laterally across the third stage. I can’t exactly say why the tension changes as the elevator goes up and down, but it almost always does.

+1 on being able to place cubes on top of layers of other cubes… Extra reach was definitely key that year.

We added extra reach mid season in 2018 by adding a mechanism that tilted our manipulator arms upward when we reached maximum extension to get the extra reach.

This design does include tensioning springs. If you look at the reply I made illustrating the cable rigging, it includes springs in each of the cable runs, unless I’m misunderstanding and you’re suggesting I place the springs elsewhere.

As for the fairlead system, I did think of that, but I decided to at least try modeling this since the lead system still introduces some angles. What I did think of so that there weren’t bearings sliding under load was doing a leadscrew on both ends and mounting the gearbox with some sort of linear slide so that it can move back and forth - but then that’s introducing a lot of complexity into the whole system. Always tradeoffs in engineering! Better not to reinvent the wheel - spools work just fine as they are. This design was meant to be a sort of practical take on an impractical idea, just for fun.

If I did a differential windlass, it was on accident. The other drum is just meant to be for retracting the elevator back down. If I’m understanding the wikipedia page correctly, it seems like it requires the two spools being connected, and currently, they are not - they both just connect to the carriage, and then the other end connects to the spool.

2 Likes

If you re-rope your elevator to have pull-down lines on each stage, then the amount of pull down rope on the winch is the same as the pull up rope. With that kind of roping, you could use a double drum traction winch that never changes rope angle as it spools in and out. You may find that less complicated that the scheme you developed. See http://www.ssg.whoi.edu/ssg/winch_wire/UNOLS_Winch_wire_Handbook/11_Traction_winch_systems.pdf Traction drums are relatively easy to make because the grooves are circumferential. It is best to drive each drum with its own motor or let one be an idler, in which case you will need twice as many grooves for the same hoist force.

Why are you moving shaft? You can use ratchet blocks. (I have never tested this before). The ratchet block could be rotated when you wind the spools.

Ropes will drag out of them with decent force on the rope or with not enough force on the rope pushing it into the pulley block. There’s also a type of sailing block that doesn’t change direction like a pulley but it holds better: a cam something.

Unrelated: thinking about it, I’m not sure which of sailing gear or robotics parts is more expensive.

3 Likes

I think you misunderstood me. (Sorry for the bad explanation. )


-Topside view
I am talking about this type of rotation. (Spool doesn’t move. )

I agree it will not hold. However, the main goal is to keep the cable at an effective angle.

It is just an idea, maybe some team will come up with a cost-friendly alternative.

I’ve been down this rabbit hole. Sailing is one of the few sports as/more expensive than FRC. Harken (Micro | Harken Marine) blocks are very common in dinghy sailing. I’ve tried with varying degrees of success to use sailing hardware in FRC. Typical sailing rope diameters are comparatively large to the typical dyneema or other lines used in our community. Even with the micro series of blocks, I’ve run into issues with them derailing.

Mostly, I’ve found that previously explored solutions within the scope of supply of our typical vendors are more cost effective and competitive. I hope someone can prove me wrong - I’d love to see some cool nautical-type rigging on a competitive robot.

1 Like