I'm working on a mechanical engineering design presentation about guni

Sleeve design change

Alright. We were having a lot of trouble trying to come up with a design for the gear shifting, since it is extremely difficult to create a device that would move the outer sleeve in discrete steps in both directions without requiring an overly elaborate apparatus.

To solve this, we changed the sleeve design a little. The new inner and outer sleeves can be seen in Figure 1.

Figure 1: New sleeve design

In the figure, the axle passes directly through the section shown. This, of course, is connected to the cranks and will be turning. Directly surrounding this are the two sleeves, which remain stationary relative to the frame. The inner-sleeve can be seen attached to the bottom bracket on the frame (I’m not sure if this is the best way to do it, but it will do for now). The grooves on the inner-sleeve have been changed to spiral along the surface of the sleeve. Slots have been added to the upper-sleeve. Therefore, when the upper sleeve is turned, the bearings will slide along the path of the lower-sleeve.

This means that the upper-sleeve no longer slides. It simply twists. This simplifies the gear-selection problem greatly. Additionally, now, provided that the outer-sleeve is held in place, the ball bearings will be locked in their location along the shaft. This reduces the chance of accidental sliding of the bearing.

Several changes need to be made to the overal design as a result, but they are minor.

Figure 2: Cross-sectional view of hub without modifications

Changes which must be made to Figure 2 are:
-The upper-sleeve must extend as far as the lower-sleeve.
-The part of the arm which slides to give the 1:1 ratio must not be pushed into place by the edge of the sleeve, but by the actual ball bearing. This is not a problem, since the grooves now extend that far
-Since the sleeve no longer has to slide, all of the space on the left hand side is suddenly freed-up. There is even the possibility of putting another planet-sun combo in the extra space, but let’s not get ahead of ourselves.

So basically, we’ve converted the gear-shifting issue into a simple matter of rotating the upper-sleeve. The amount of the rotation per gear will depend entirely on the pitch of the grooves in the lower-sleeve. There is ample room between the frame and the edge of the hub for some device to control the rotation. It basically just comes down to preference now. What would people like to see controlling the rotation? Some kind of ratchet? Bike cable and grip shifter?

As an aside: All the comments about gear ratios are duly noted. I don’t know if we’ll have time to make the changes for this presentation, but the lower step ups might be a good idea.

Eric shows you a really nice gear spread and his gear diameters are consistent. Then it requires realistic pitch selection to get enough tooth cross section to avoid shear. There are a couple of items in your design which raise questions.

Your sliding ball scheme is really cool. You can just see those things move along parallel to the axis of the axle, engaging the sun gear detents as the outer sleeve twists. If the twisting sleeve can snap the balls into the sun gear detents with hand pressure, however, how hard would it be to pop them back out with the huge force your legs can apply?

Florian’s scheme for attaching the sun gear is really clever. He would bring the entire sleeve out, machine a boss in the sleeve that mocks a 42mm bearing outer race, press a bearing smaller than 42mm into the sleeve boss and grip the boss with the bearing clamp. Very clean and cylindrical.

Another problem I see is the coaxial assembly comprised of the axle and two sleeves. Running those three with bushings would add a great deal of rotating friction to the assembly. I don’t see any provision for ball bearings between those rotating surfaces. The space between the sleeves is unimportant and could easily just be a sintered, oil impregnated bronze bushing surface. The space between the inner sleeve and the axle is a different story.

For animals who have benefited from millions of years of evolution to achieve the convenience of opposable thumbs, get the shifter away from the foot and into the hand. Your shift lever is right there outside the hub and not rotating. A cable running up the frame to a shift lever would be a huge boon to unicycling. Will it still work when shifting to 1:1 mode though? Then the sleeves have to rotate within one another and the ball bearing problem resurfaces. There’s not much real estate in that hub, is there?

Hey, that’s another question. How do you plan to manufacture the hub which looks like it has an integral ring gear? I think Sturmey Archer had some models with ring gears integral to the hub.

Looks good, there definitely is room for another gear in there…

For the shifting I would like to see a single cable shifter compatible with front shifters on bikes. That way people could use grip shifts, trigger shifters or bar end shifters.

Hmm…I see what you’re saying. I completely missed that.
How about we switch spiral and the straight slots. If the lower-sleeve, which is stationary relative to the frame, has longitudinal grooves instead, it will take the full force of the sun gears, since they wil apply a force perpendicular to the groove.

The upper-sleeve could then have the spiral groove. I guess that might cause a problem for shifting gear under load though. I’ll think about that one. Thanks for pointing it out.

I figured movement between the sleeves is almost negligible, since it is always very low speed, and only occurs when changing gears.

The lower-sleeve is securely fastened to the frame (maybe not in the picture I drew, but ideally). There should be a way to provide enough clearance that it never touches the axle. It shouldn’t ever experience any uneven forces causing it to bend and touch the axle. I don’t know if it’s wishful thinking, but I think a sufficient gap is all it needs…

Hadn’t thought that far ahead…:slight_smile:

Why do the balls have to pop into the sun gears? Why not just let them slide in axially and have the torque hold them in place circumferentially? Let’s say you’re still using balls instead of some intricately shaped indexing block. Let’s say there is a space between the two sun gears equal to the diameter of one of the balls. That way the balls leave the engaged sun gear and travel down the slots to the free wheel space. No parts have been elastically stressed. The hub freewheels until the sleeve slots line the balls up with the slots in the unengaged sun gear. That way the balls can’t lock the two sun gears together. If this is done quickly enough the hub doesn’t freewheel too long.

Seems like the cable-shifter crowd is winning. I shall try to convince my teammates.

If this could be done with confidence that the balls won’t slip out while in gear, then that would be great. There’s more than enough room for extra clearance in the hub now (provided I sort out the kinks with the twisting sleeves).

I agree except for the 1:1 case. Somehow both sun gears have to be disengaged from the frame and locked to the hub. What happens with the two sleeves in this case?

I’m afraid I think that’s wishful thinking. I agree that a even a small radial gap of 0.002" (0.05mm) to 0.005" (0.12mm) is adequate spacing but in order to maintain it the sleeves and axle all must be closely indexed in radius to the hub central axis. If the either the axle or the inner sleeve is free to float as a cantilever through part of the assembly it’s not going to take much force to bend it through the width of that gap and scrape against the other surface. I could be missing something entirely here and I hope I am.

They should be stationary. There might be some wear going on at the moment though…the way the balls interact with that moveable peice of the planet arm has been redesigned many times…

Yeah, that’s pretty much what I was hoping for. I suppose we could squeeze in a set of bearings at the end of the sleeves. Then they would be supported from both ends. It would require some re-arranging however.

That’s how it looked to me, Glen. The sun gears are completely disengaged, which I had forgotten, but the balls are on non-rotating sleeves and applying a force to a rotating planet cage, scraping against it.

Man, I hope this works. Are you guys going to build this thing? There are probably a couple of guys at Moab right now who can provide constructive comments. I pretty sure Jason is there and JC might be there, too. Roger Davies, Joe Rowing, and a couple of the other engineering guys in the UK would be keen on this, too. They might be at BMW right now. Joe hasn’t been active in the fora for a while, though.

Far more important, I’m going to eat chocolate ice cream with chocolate syrup and watch a movie with my wife and daughter now.

You know…my design used to work perfectly until you started noticing things…:slight_smile:

Not in the short term, that’s for sure. I don’t have the time or the resources. In 4th year, however, we’re supposed to choose a design project to persue.

Talk to ya later!

Close to the end of their axial travel the balls could disengage from the inner sun gear and begin to push the inner race of a ball bearing which slides axially along the outer slide. The outer race of the ball bearing could push against a mechanism on the planet cage which engages the hub or ring. Further axial travel of the balls would cause the mechanism to lock the planet cage to the ring gear or hub. I don’t immediately see how this can be reliably disengaged unless the mechanism attached to the planet cage is preloaded with some kind of spring, probably a set of small leaf springs.

I’ve never used ball bearings designed to handle axial rather than radial loads but you clever ME guys do it all the time. As far as I know the races are just at different angles.

Hmm…that’s a real interesting concept.

I have a ton of studying to do today, but at some point I might take a break and whip up a few sketches of the latest iteration of this design.

Actually, something you might be more familiar with, with your experience…what is a good width for the gears? I just ballparked it at 1cm during the initial design for lack of a more educated guess.

Also, regarding sun gear design. From what I understand, you’re suggesting that the gears be ridged along their inner-circumference (in order to lock the gears), but not along their width (to allow easy entry and exit of the balls)? Is that still with the twist shifter rather than the slide shifter? If I got that right, that should work, now that the lateral position of the balls is fixed by the sleeves. One consideration for any method is that in order to reduce slop, the balls will have to have a very tight fit. I realized last night that with the low-energy-detent method, there could be a lot of movement within the detent with changes in crank torque.

The gears I use are 1 cm wide.

No. I picture the slots on the inner circumference of the gears running axially. The troughs underneath these slots are the ones on the sleeve with the straight slots, hopefully the outer sleeve. The inner sleeve has the pithed slot. With no rotation, the gear slots and straight sleeve slots can all be lined up and you can look right through them. The balls can slide down these channels but will remain in whatever location the pitched slot of the inner sleeve allows them. Since the outer sleeve has the straight slot, the engagement points of the balls are right at the diameter. There are almost no tangential forces on the balls trying to push them radially as a result.

I’ll try to sketch it.

Here’s a sketch of the gear, inner sleeve, outer sleeve, and ball set. I showed both views in cross section hoping to clarify it.

Shift.pdf (8.68 KB)

Here it is attached as a jpg.


This kind of shows the two sleeves in a chaotic sort of way.


The original reason for trying to avoid having a gap between the sun gears was that we were trying to reduce the width of the entire system. I see what you mean about making the sun-sleeve interface occur that the diameter of the ball though. That would really increase the amount of torque the system could handle.

In order to have the straight slots on the outer sleeve, it’s not too difficult to find a way to fix the outer-sleeve directly to the frame, and attach the twist-shift mechanism to the lower sleeve. Either way, I’m still not quite sure how well shifting under load will work.

I’m working on some drawings.


I kind of ran out of time to work on new things, since the presentation is tommorrow, so I threw together a semi-up-to-date drawing to stay consistent with some of the other sketches we’d already made.

I attached the outer-sleeve to the outer race of the bearing that the frame clamps to. That way the outer-sleeve can be the one that has straight slots, and also withstand the torque applied by the sun gears. There’s a small opening on the top of this sleeve that can be seen in between the bearing and the left side of the hub. Some mechanism could control the rotation of the lower sleeve, either a boss that could be rotated, or more gears. Not sure if that’s the absolute best design, but it’s good enough for the presentation.

I kept the curved sun-gear design for now (and even grossly over-emphasized it for explanation purposes. I’m aware that as drawn, they would never actually manage to slide from one gear to the next). I also left it with 1:1, 1:1.5, 1:2.25 to stay consistent with some of the other things we’d already written. It’s not a perfect drawing, and hasn’t been hatched properly. There’s room to sneak in a bearing or two near the sliding portion of the planet arm, but I left it out to avoid cluttering things. I also left out the spirals on the lower arm, but trust me, they’re there. The smaller of the two sizes of bearing that I included was arbitrarily 1cm x 1cm. A lot of my measurements were either educated guesses, or based on measurements of my own hub, but for an idea of size, the absolute OD of the hub around the axle as seen in the picture is 7.5cm. This would increase slightly if geared for lower steps.

Nice! I like to see the development going around these multiple speed unis.
Have fun with the presentation.