Vex Tri-star wheel -- A work in progress

VEX V5 General Discussion General Forum
1

Heyo, VEX forum!

All the holonomic modules and mecanum wheels and whatnot being posted here in the past couple months inspired me to finally do some serious designing on something I’ve always been fascinated by: Tri-star wheels. I’ve built several out of LEGO mindstorms, and have CAD drawings somewhere of an FRC-scale wheel, but creating one with VEX posed a real challenge.

I wanted to avoid simply designing laser cut lexan sideplates, as I feel that this is a valuable resource that should not be spent entirely on the drive. A set of 8 plates would burn up the sheet pretty quickly. Finding a VRC-legal way to create a solid, compact tri-star arraignement wasn’t easy, but I recently hit two breakthroughs: VEXplorer wheels, and 1/4th pitch metal.

The VEXplorer wheel is the only VEX piece with a 6 hole bolt pattern. These wheels could be cut down to form the plate shown on the right, which formed the basis of my design.

http://img16.imageshack.us/img16/7360/vexplorer.png

From there, a metal structure was added on, providing support for all appropriate bearings and gears

http://img62.imageshack.us/img62/8163/plateo.png

Two of these sideplates, linked together with some standoffs, along with some gears and wheels in between, form a tri-star wheel!

http://img828.imageshack.us/img828/5701/gearing.png

http://img832.imageshack.us/img832/6057/isofront.png

http://img23.imageshack.us/img23/4402/topsot.png

(If the bearings look strange, its because they were custom modeled using the photo on vexrobotics.com as the only reference…this part appears to be missing from VexCAD, and I don’t have any 1/4th pitch metal at the moment)

Thoughts? Questions? Comments?

2

A couple more angles of the final product

http://img541.imageshack.us/img541/5005/isogear.png

http://img825.imageshack.us/img825/52/frontdp.png

3

Fascinating! I have waited a long time for someone to do something with planetary gear systems!

I would really like to see this built as well as implemented on a robot!

4

Because there is soooo much air (and so little metal) in them, the 1/4" pitch metal parts bend very easily. Without some sort of extra bracing, I worry that the structure will bend or collapse very easily under side-loads.

Even using using the typical Vex metal (1/2" pitch holes) I would worry about this.

Noticing the six hole pattern in the Vexplorer wheel was a good catch, but without a method for adding more rigidity to the entire structure I’m not optimistic it is practical yet.

Long ago, when I was looking for a way to make some triangles I used the holes in the tips of 45 degree gussets to link together the tips of short C-Channels. The result was a triangle, but it would deform easily when the joints between the gussets and the channels would slip (a single screw made those connections).

I dunno a good answer for either of us - Let’s keep looking.

Perhaps the answer would be to make (out of Polycarb) a set of 3 straight/bent links that will form the same hole pattern that you are getting from the sliced up wheel in your current design, or a set of multi-hole triangular gussets that will replace my 45 degree links? Triangular gussets or straight links will not consume as much Polycarb surface area as open circles will.

Blake

5

This is awesome. I didn’t even realize the VEXplorer wheel was legal. Silly silly me. Making a plate like that could be useful for other things too…

But anyways how thick do you have that VEXplorer plate being currently?

PS- It was good to stay away from the lexan anyway, because it can only be 1/16" thick, and I could forsee a lot of bending/cracking with lexan plates. Nice discovery of the VEXplorer wheels.

6

This link has some good background on tri-star wheels:
http://orionrobots.co.uk/Tri+Star+Wheels

It mentions needing some slipage to get the walking mode working.
Something about needing a differential between the subframe rotation and the central axle rotation; but I didn’t understand it.
Does that match your experience with lego models?

The cantilever mounted 12 tooth gears will be wobbly even if you add a locking collar to the other side to retain it in place.

If I understand the theory of operation correctly,
the current gear ratio is 1:1:2.75" in roll mode and ~1:6" in walk mode.
Wouldn’t it be better if walk mode had equal or more torque than roll mode, rather than 2x less?

Quad-star wheels, may be easier to prototype and work out issues with the rotating subframe and gear ratios.

Comparisons to planetary gear system are interesting:
In roll-mode, this system has 2 planets in series in a fixed carrier, and the ground is the ring gear.
The idler gear shown here is needed to make the wheels rotate in same direction as the main shaft. If you used sprockets and chain instead, you wouldn’t need the idler gear; it would probably take a sprocket on one side for 2 wheels, and a sprocket on the other side for the remaining wheel (or 2 for quad-star).

7

First, very cool. Congratulations.

Second, I’m pretty sure that a wheel assembly capable of traversing rough terrain isn’t going to be useful in the VRC this year, so why not free yourself from the game rules and build your assembly using whatever you want? A piece of 1/4" polycarbonate might be just the thing to build your side plates.

Thank you very much for sharing this!

8

I agree with all that was said. Nice find on the vexplorer wheels. This is certainly an interesting design. :slight_smile:

Could these possibly have a gateway application as being used as an intake roller? Rather than wheels? Just a thought.

9

http://img16.imageshack.us/img16/7360/vexplorer.png
So I wanted to look more into this to see if the part you made could be used for a planetary gearbox. Unfortunately, it isn’t big enough (wide enough) to fit any ratio other than a 1:1, which at that point is redundant. I just got so excited when I saw that: it looked so similar to the parts from the andymark planetary gearboxes, I was hoping it would work. Still, a very cool concept you found.

My only other idea with the vexplorer wheels is a mecanum wheel. After seeing how green egg took the 60-tooth high strength gears and milled them to make mecanums, the vexplorer wheels seem like they would work even better for an application like that.

10

https://vexforum.com/gallery/showimage.php?i=1856&c=searchresults&searchid=595804

I figured you might find this helpful. It isn’t as clean as your design in my opinion but its the same concept.

~DK

11

I thought I remembered seeing one in the VEX Gallery somewhere.

Good luck on designing this, I’ve built some out of LEGO and built a robot using 4 of them before, as well.

~Jordan

12

heres a lego youtube vid

these things are AWESOME!!
perfect off season project :slight_smile:
or for next year if there is more terrain

13

The tricky thing about tri-star wheels, so I’ve been told, is that the input has to be fed with a bigger gear than the output. I.e. the outer gears must be smaller than the center. I don’t know why, but I’ve heard this many times.

14

This would give the whole wheel more torque than each individual wheel, so it “locks up” more readily and will climb easier than the 1:1 that is shown here, especially if the surface is slippery. However, you do lose torque on the individual wheels with smaller gears on the outside, which could make it “lock up” more often than you want it to, or on surfaces that aren’t that bumpy.
The smaller the outer gears, the faster the individual wheels will go with respect to the whole thing.
It’s all a torque problem.

15

So here’s an idea…

what if someone took this tri-star wheel, and put omnis on it? and then turned it into say, a kiwi drive, or a x-holo drive?

hmmmmmmm…

16

that is realy cool! i would like to see that in action

17

How would it climb, then? To achieve omnidirectional movement, I think it would require mecanum wheels.

18

That would be scary

19

If only the VEX game required robots to be able to surmount obstacles… :smiley:

20

Thats 3 people now that think the Sun gear needs to be larger than the wheel gear, although the lego video doesn’t show it to be much larger. You may have to gear-down the motor to drive the central axle, so you can gear it up again at the roll-mode wheels.

This is an old picture, without any followup. Due to lack of idler gears, its only good for show, without full functionality of the roll+walk wheel shown in the lego video.

Other notes:

  • Planetary gearbox can have 1…N planet gears. 3 is common and effective, but not magical. 4 would work fine.
  • Planetary, tri-wheel, and differential all have some interesting commonalities.
  • Sprocket& chain setup doesn’t need idler gears to get correct rotation, and axle spacing is more flexible. Some chain slop might be beneficial.
  • Getting 120 degree angles is not critical! correct radius is, but you can use radial vex metal strap to help position the axles, and the wheel frame just as a brace.
  • If you are willing to cut up vexwheels, why not other sorts of wheels? The original 4" and 5" vex wheels are larger, and you can cut the 3 holes in their rims anywhere you want. Or just use 1/4" lexan as suggested by Rick, at least for prototyping.

Here are some other possible ratios (Sun,idler,wheel gears) and comments:
– 24+12+12 : 0.5" hole pitch is 1.5h+1h : total sun-wheel axle distance 2.5 h = 1.25" ; this is the smallest vex gear size that has Sun gear > wheel gear
– 36+12+12 : 2h+1h =3h = 1.5" radius of subframe
– 36+36+12 : 3h+2h = 5h = 2.5" radius = try 5" wheel as subframe?
– 24+36+12 : 2.5h+2h = 4.5h = 2.25" radius = 5" wheel? 2:1 speedup.
– 36+24+12 : same as above, except 3:1 speed up
– 36+12+24 : same as above, except 3:2 speed up
– 36+60+12 : 4h + 3h = 7h = 3.5" radius subframe + 1.35" small wheel radius = nearly 10" total wheel diameter in walk mode,
---- 60+12 is minimum idler-wheel gear needed to avoid cantilevered planetary gear

It might be worth considering to cut the flaps off an intake roller and use them as wheels, or find other ways to get smaller wheel sizes, to prevent the whole wheel from becoming too big.