Expandable Gear CVT: Michael Warren Reed's Project

[MichaelWReed]

Hi there,

I started working on an idea for an expandable gear CVT about 3 years ago, while out of work for a few months.
I had some exciting progress on an idea that I think might be worth pursuing more, but I now have 2 kids, 2 jobs, and no time. I have brought up the project idea with a few friends, but they didn't have the time either, or they wasn't interesting enough to them to strain their brain :)
I originally thought of the idea while riding my single speed bicycle, which I like to due the high toque to speed efficiency, but I still wanted to be able to adjust the gear ratio somehow for going up or down hills, of course.
I really like the NuVinci CVT, created by Fallbrook Technologies, just north of me in Cedar Park, TX, I live in Austin. But I suppose I like my idea better, if it works, due to it having a higher speed potential, and for it potentially being the first (I believe) solely mechanically automatic transmission.
Any feedback and/or collaboration on moving forward with this project would be appreciated.

Please see the attached MS Word .docx file, and concept images.

Thank you,

-Michael Warren Reed

 

[berkeman]

Cool. It looks like you've done some prototyping -- have you fit a prototype on your bicycle?

 

[louie13]

This is very interesting, have you ever considered fitting the mechanism inside of a hub shell where the rear cog would be the one that expands/contracts?

 

[MichaelWReed]

@berkeman - unfortunately I haven't yet made any prototypes, having some rapid-prototypes from a 3D CAD model would be very helpful, and I have Autocad software, I just haven't had the time to sit down and make the models.

@louie13 - thank you, I'm thinking the whole setup would probably work best in an enclosure of some type, to protect the large number of moving mechanical parts and to make the setup more stable. The problem with only having one cog expand/contract is that the chain isn't elastic, and so wouldn't allow for expansion, or would fall off under contraction, unless springs are used like in typical multi-speed transmissions. When 2 cogs are used, inversely correlated, then the chain can maintain the same fixed size, maintaining that nice high torque conversion I'm looking for.

 

[louie13]

@berkeman - unfortunately I haven't yet made any prototypes, having some rapid-prototypes from a 3D CAD model would be very helpful, and I have Autocad software, I just haven't had the time to sit down and make the models.

@louie13 - thank you, I'm thinking the whole setup would probably work best in an enclosure of some type, to protect the large number of moving mechanical parts and to make the setup more stable. The problem with only having one cog expand/contract is that the chain isn't elastic, and so wouldn't allow for expansion, or would fall off under contraction, unless springs are used like in typical multi-speed transmissions. When 2 cogs are used, inversely correlated, then the chain can maintain the same fixed size, maintaining that nice high torque conversion I'm looking for.

Ah, this makes more sense, some sort of chain tensioner would be necessary if it were to be implemented into a hub shell.

 

[MichaelWReed]

Ok, so I was chatting with a friend about my "solely mechanical automatic transmission" claim, and I realized that that might not be the best wording.
What I am meaning to claim is this would be the first non-electrical, non-hydraulic automatic transmission; in that the shifting is being solely affected by the spring in between the crank and the crank plate, balanced against the torque applied to the crank.
I also am looking at the write-up, and realizing it's a big old mess, and doesn't flow very logically, I will work on that, and a sketch of the full system with both front and rear expandable gears.

 

[MichaelWReed]

Here is my updated write-up:

Expandable Gear CVT

Abstract:
Single front and rear gears that both expand and contract (using looped scissor-like radial expansion segments, Hoberman), inversely correlated (as one expands the other contracts), duplicating the effect of popular cone shaped variable transmissions, but with the high torque to speed conversion efficiency of a single front and rear gear setup.

Description:

- Expandable Gears:

Gear Contraction (front gear, on crank):
This will happen when the expanding circle (Hoberman) gear is held rotationally still, and the gear core in rotated. When the gear core is rotated, the curved expansion arms will rotate and move from a horizontal position to a vertical position; pulling in the expansion arms; contacting the expandable gear.
Rotating the gear cores in relation to the expanding gears will be accomplished with a clutch plate attached to a disc hard attached to the crank axis. This clutch plate, along with the crank axis, is moved inwards to connect with the gear core by way of an inner crank plate moving along screw threads into an outer crank plate.

Gear Expansion (front gear, on crank):
This will happen when the gear core is rotated in reverse. This will be accomplished with a planetary gear setup on the crank axis, on the opposite side of the gear core than the expansion clutch. The outer ring gear will be hard attached to a plate, hard attached to the crank axis. The inner sun gear will be loosely attached (oil or bearings) to the crank axis, and would have a clutch plate attached to it, to make connection with the other side of the gear core. The middle planet gears will spin on pins hard attached to the plate that the outer ring gear is hard attached to, and would be held in place with pin heads or by enclosing the planetary gear setup (minus the sun gear clutch connecting surface).

Gear Expansion and Contraction (rear gear, on rear wheel axle):
The rear expandable circle (Hoberman) gear will need to be inversely correlated to the front expandable circle gear, and will need to be perfectly synchronized. This will be accomplished by hard attaching small gears to the side of each gear core (front and rear), just outside of where the clutches connect, and connecting these gears with a standard bike chain. The rear expandable circle gear will be flipped around, so that the rotational movement of the front gear core causing expansion will cause contraction in the rear expandable circle gear.

- Drive Chain:
The inner most rotation points of the expanding circle (Hoberman) have the greatest ratio of expansion, and so the drive chain would need to be expanded/contacted based on their movement. For the drive chain to be expanded/contacted along with the circle, the rotation points would need to have expanded pins (out towards the crank plate), long enough to support the width of the drive chain, with freely spinning cylinders on which the drive chain would rest and move along like a conveyer belt.
The drive chain will have the gear "teeth", which will protrude horizontally, towards and connecting with the crank plate and the rear axle plate, which will have radial grooves that the teeth connect with. The teeth would connect in unison with the rotation of the crank plate on one half of it, but the issue would be the other half, where the chain and teeth would cause friction against the rotating crank; moving somewhat perpendicular to the rotational direction of the plate. To solve this, the "teeth" need to be able to swivel horizontally out, and then back inside the chain; so that they can come in frictional contact with the crank plate only on the one half where it is helpful. This could be done with chain "guides", one that will brush the back of the teeth causing them to open up outward just after the half way point, and then another guide to push them back into place at the other half point. This would be the same connection type on the rear axle plate.

- Automatic Transmission
The automatic transmission will be accomplished with a spring (*this could also be accomplished by filling the space between the two crank plates with compressible fluid or gas, which would be compressed when the inner plate screws inward, and would “spring” back when the torque pressure goes down). The spring will automatically expand or contract the gears, based on the torque needed to move the bicycle. This will cause the level of torque needed to move the bicycle to even out to an average-constant.

The crank arm will attach to a smaller, inset crank plate, nested in the larger crank plate, the spring connecting the two plates together, vertically. When torque is applied, the inner crank plate will start winding the spring. If low toque is needed to rotate the rear wheel axle, then the spring will wind a smaller amount before the outer crank plate/drive belt/rear axle starts to rotate, if more torque is needed (i.e. going up hill), then the spring will wind up more beforehand.

The smaller crank plate will have a hard connected protrusion (lock pin) on its vertical outer edge, which will lock it against the vertical inner edge of the larger crank plate, at the starting (0% torque) position. This will prevent the spring from pulling the crank plates into a negative torque position.

The relative difference in rotation of the smaller and larger crank plates will be translates into movement of the crank axis in and out of the bicycle frame. This will be accomplished by having screw threads connecting the small and large crank plates. As increased torque causes more relative difference in rotation between the crank plates, the smaller crank plate will begin to rotate on screw threads between its outer edge and the inner edge of the large crank plate; moving the small crank plate inward. As the small crank plate will be hard connected to the crank axis; as the crank plate moves inward, so too does the crank axis.

As torque starts at a low percentage, the planetary gear clutch will start off being connected, causing front expandable gear expansion. As torque increases to mid-range the crank axis will move and disengage the inner planetary gear core clutch, stopping expandable gear expansion and maintaining it. As torque increases to a high level, the crank axis will move in enough to engage the outer front gear core clutch, causing front expandable gear contraction.

Torque adjustment would be accomplished by “winding” or unwinding the spring.

- Manual Transmission

Two cables will be connected to the MC. One cable will rotate the MC counter-clockwise (looking at the crank side of the bicycle); engaging the GCDB (gear expansion). The other cable will rotate the MC clockwise; disengaging the brake, and then engaging the CPGSC (gear contraction). The torque spring, smaller crank plate, small disc plates and TAC connected to the MC are not needed in this version.- Stability:
The entire setup would be likely to contort under the torque pressure.
The pieces making up the expanding circular gear will need to be I-Beam shaped hinges; to avoid side to side bending.
Due to the issues of dirt and other environmental factors affecting the complex joints and pins, the entire setup will be enclosed in a thin casing (metal, clear resin?), exposed only at the rear wheel axis, this should also assist in minimizing the contortion.

Torque Adjustment Cylinder (*originally included but... needed??)

 

[MichaelWReed]

Drive Train -
So... I just realized a major issue with the drive chain. In order to allow for expansion/contraction, the chain needs to have one-way friction along the contact points; to allow the contact points to move back along the belt. BUT, I wasn't thinking about how the rear gear contacts would need to be in reverse! Instead of being pushed, they are doing the pushing.
I had originally tried out the idea of an elastic or spring connected drive chain, that could flex to support the expansion, but the issue is that these points would also flex while applying torque to the belt to move the crank, so you would want it pretty tight, but then you need it to be pretty loose so that it doesn't take a lot of force to expand the contact points during expansion.
So, I suppose I might need a double sided chain or belt, that has opposite direction one-way friction connection points on either side. The issue there is that then the rear gear connection would have to be on the inside of the contact points on the gear, which would be difficult, or... the chain could be twisted? Twice, one between each change over between gears.
Or, there are one-way hinged-door type chain connectors (with small springs inside to keep them up until expansion occurred) having a sliding switch on the side of the chain so that when slid forward or back they would allow or prevent the door from swinging one way or the other, with some mechanism to switch the slide lock on each, two times, as they pass between each gear.

 

[Baluncore]

Your project is interesting because it is too complex for most human minds. The length of the cranks are fixed by the geometry of the human body. You are playing with the chain gear ratios.
But what if the rear wheel could change it's rolling circumference depending on torque? Developments in solid tyres with elastic hubs may make some new wheel technology the answer.
Think of a clock spring. Maybe the rear wheel would be a one turn spiral when not doing work, but would then form say a three turn spiral when the full weight of the rider was being applied. The outside diameter of the wheel would change by a factor of three. That might be easier to engineer than a chain or stepped belt drive with a variable ratio.

 

[MichaelWReed]

@Baluncore - Thank you for saying that ego booster, but what I'm taking from this is that I really need more pictures, 3D models, animations of the mechanics, etc in order to communicate the concept, and I agree.
I'm trying to follow your rear wheel diameter change (based on torque applied to the clock spring setup), but wouldn't the rear of the bike be raised and lowered during this change? That might be a strange riding experience.
So this spring would be set to an expanded state, so that when little torque is applied the spring/wheel is largest, then when more torque is applied, the spring/wheel contracts.
My worry with this idea is that the strength of the spring needed to both expand the wheel as well as hold up the person riding the bike would be too great to be overcome by pedaling, but I'm not sure.
I can see that when an increase in torque is needed to maintain the same speed while riding, the torque would transfer from bike speed into contracting the spring, until the torque is no longer increasing, or the spring/wheel reaches its smallest size, and then vice versa when the torque needed is less.
This is also a fun idea to try, I'm just not sure about the rear of the bike raising and lowering while riding, unless that's not what you meant.

 

[Baluncore]

That might be a strange riding experience.

It certainly would, especially when climbing hills on your back, or descending steep hills over the handlebars. Overcoming that diameter change would be another challenge.

You are in good company. There have been many dreams of continuously variable ratio toothed gear systems. Only one or two have survived to hide in narrow niches in the market place. Search for CVT, continuously variable gear, or Positively Engaged Mechanical CVT.
An example of a recurring dream is here; http://vmt-tech.blogspot.com.au/

 

[MichaelWReed]

I'm intrigued by the VMT Universal Transmission, but they seem to be not wanting to share the details on how it works, other than that there's moon gear that's expanded and contracted, somehow.. do you know how it works?

 

[Baluncore]

It probably does not work. They are willing to charge you a fee to demonstrate that it is not applicable in your case.
Patents are listed here ...
http://patents.justia.com/assignee/vmt-technologies-llc
Get the pattent application number, prefix it with US and google ...
http://www.google.com/patents/US20110306457