CVT (Continuously Variable Transmision)

[Solcar]

Has anyone done experimentation with a CVT (Continuously Variable Transmission)? I've been looking for a small, light weight one other than a wheel hub type (pretty big, costly, a bit heavy?) to try out on an ebike. We have talked about using a pressure roller that has different diameters along its length that would be moved from side to side, but a potential implementation of that looks less than ideal, for example, there would be more time needed to get the roller to move left or right if the roller uses that diameter stepping than to just widen or narrow v-belt pulleys. Or if the roller were to be conical, the contact patch on the tire wouldn't move follow a curved motion, increasing tire wear. The type that moves the roller left to right is hard to build mounts for.

I want something specifically for low power that has good pull over a wide range of speeds, especially good for efficient low speed efficiency or on hills.

[Dauntless]

Heavens, the threads on that around here. Have you searched NuVinci? I for one am all for them.

http://www.ebay.com/itm/49cc-X1-X2-Ninj ... 51&vxp=mtr

[Hillhater]

There are several threads on multi ratio /CV friction drives if you keep searching, but one conclusion was that you really do not need a cv/multi ratio drive with electric power. Just 2-3 well chosen ratios works well for efficient power use.

Heavens, the threads on that around here. Have you searched NuVinci?

... I suspect he is aware of NuVinchi...

I've been looking for a small, light weight one other than a wheel hub type (pretty big, costly, a bit heavy?)

[John in CR]

I ran a Comet Torq-a-verter belt type CVT on a low powered emoto that worked ok, but they're pretty big and heavy. I remember someone running a variator from a scooter on an ebike and another on an emoped, and those are similar takes on the same theme. The trick is to tune them for a match with our lower rpm higher torque electric motors unless you're going to use a higher rpm RC motor which are closer in rpm to what the variator for gasser scooters want.

[Solcar]

There are several threads on multi ratio /CV friction drives if you keep searching, but one conclusion was that you really do not need a cv/multi ratio drive with electric power. Just 2-3 well chosen ratios works well for efficient power use.

I indeed remember that and though just two or three ratios can do well, some reasons why I want to avoid doing it with a gearbox with selectable ratios is the complexity, like with the idea of reversing the motor to switch gears, and the smoothness of the CVT, especially, for example, when going from drive to braking. Say I'm going along at 15mph and begin to descend a dangerous hill, the ratio can be adjusted to very high gearing and then the throttle let off, going from propulsion to drag, but since the ratio is at a high setting, the drag is not too much and the 15mph can be maintained. Other times it can just do the job of the brake generally.

I think some smart folks out there are probably capable of doing it simply and effectively, so I wanted to bring up the topic again, even if those folks don't usually post messages.

I've now added the belt between the pulleys in that drawing I've edited in the first post.

... I suspect he is aware of NuVinchi...

Indeed, thanks.

[Solcar]

I ran a Comet Torq-a-verter belt type CVT on a low powered emoto that worked ok, but they're pretty big and heavy. I remember someone running a variator from a scooter on an ebike and another on an emoped, and those are similar takes on the same theme. The trick is to tune them for a match with our lower rpm higher torque electric motors unless you're going to use a higher rpm RC motor which are closer in rpm to what the variator for gasser scooters want.

I saw a video of one operating on a drill press, and it looked pretty cool. The big and heavy aspect is the thing that I'd like to see get overcome. Indeed, I use a small motor.

[Solcar]

Heavens, the threads on that around here. Have you searched NuVinci? I for one am all for them.

http://www.ebay.com/itm/49cc-X1-X2-Ninj ... 51&vxp=mtr

Quite an interesting gearbox.

[Harold in CR]

There is a binch of scooter CVT,s on Ebay. One was $25.00 that I almost bought.

[Solcar]

If I understand, those operate by centrifugal force. Later I plan to show a diagram of the thing I want to try from a side view. I'd wish to control the ratio of mine manually via a cable or something like that. Yet on another project, I might try one of the scooter types.

[John in CR]

If changing direction of rotation for your two gears is appealing, then retro-direct is what you want. The compromise is that the bike will no longer roll backward. It's elegantly simple requiring just 2 extra sprockets and freewheels.

[Solcar]

The set-up I would like to switch my ebike to would pretty much involve having a wide range of gearing to choose from, like the through-the-gears set-up I have now except it would be much lighter weight. I ride slowly, so low power with wide gearing does well for me (I take my time on back ways when feasible). I also hope it would be able to slow the bike down on a few dangerous hills around here without having to ride the brakes, and give good push up some steep hills. I think that most people don't need much if any variable gearing options because with a lot of power it doesn't seem to matter as much.

[whatever]

can anyone point towards ebay cvt that you can use cable to adjust ratio?
the previous ebay link is the ratio controlled by cable? or some sort of auto adjust??

I'm wondering about the efficiency of these things to?

[gwhy!]

can anyone point towards ebay cvt that you can use cable to adjust ratio?
the previous ebay link is the ratio controlled by cable? or some sort of auto adjust??

I'm wondering about the efficiency of these things to?

I came to the conclusion when I converted a 50cc auto scoot ( with CVT ) to electric that on the face of it the loses through a cvt is pretty bad ( I was loosing around 200-300 watts through mine but i'm sure this could be brought down to more like 100-200 watts if everything is engineered to a higher spec ) BUT you will gain efficiency due to the nature of what the cvt does ( its always in the correct gear for the load if its setup correctly ) so they balance each other out, I think if you have a cable to adjust the ratios this efficiency gain will be lost

[John in CR]

The set-up I would like to switch my ebike to would pretty much involve having a wide range of gearing to choose from, like the through-the-gears set-up I have now except it would be much lighter weight. I ride slowly, so low power with wide gearing does well for me (I take my time on back ways when feasible). I also hope it would be able to slow the bike down on a few dangerous hills around here without having to ride the brakes, and give good push up some steep hills. I think that most people don't need much if any variable gearing options because with a lot of power it doesn't seem to matter as much.

Motor cogging isn't enough to slow you going down hills regardless of gearing. You need regen or plug braking for that, so without some engagement/disengagement mechanism you have to forget about freewheeling, ie motor drag for pedal only riding. A nice soft regen setting is great going down hills and keeps the brakes nice and fresh, so they're ready for emergency stopping at any time. Greatly reduced brake maintenance is what I like most about regen, with the added safety going down mountains a close second.

The only rig I know of that will do what you want is the 2 speed high efficiency hubmotor I'm testing. In wheel you'd be stuck running a scooter wheel, but as a mid-drive use any size wheel. The power line would be dead simple using 2 fixed sprockets (pulleys) and a chain or belt, so you'd still have regen. In low speed it will climb anything unless you set the gearing too high (over about 30mph top speed in low and 50-55mph in high speed at 74V). Despite being capable of multi-kw it's really compact at less than 7" OD after trimming off the 4 rim tabs. It's far more efficient than other hubmotors, which means no overheating on hills and 10-20% more range depending how you ride. It's by far the best motor I've tested, probably because it's a Japanese design instead of Chinese. It's so good that I'm considering building a business around it.

John

[Solcar]

Your project is really cool, especially for its simplicity. I'm planning on something that is just a few hundred watts tops. I hope to set it up as friction drive having something like a 2" diameter roller that can be driven from a small continuously variable reduction that is cable controlled, like how the old ten speed mounted derailer lever was mounted on handlebar stems.

[motomech]

CVT's are continuously variable between their fixed low and high ratios
The way they they work is a function of three things,

1]drive pulley rpm. The outside half of the drive pulley[called a sheave, the inner sheaves are fixed, the outers are moving] contains 3 to 6 roller and ramp assemblies[weights are sometimes used instead of rollers]. As the drive pulley rpm increases, centrifugal force moves the rollers outward, spreading the sheaves apart and allowing the belt to ride lower, moving the overall ratio downward[numerically higher]. The number and weight of the rollers can be changed as a tuning measure.

but there is more,

2]The driven pulley sheaves are pressed together by a stout spring on the outside of the movable sheave. Spring tension can also be altered to adjust the rate that the CVT moves though it's range.

Lastly

3]The load on the rear wheel[ie., the driven pulley].

In general, the drive pulley influences the initial rate of change, while the driven pulley goverens the rate the latter.

I really don't see how any type of manual adjustment of parameters could be applied with the exception of a system to regulate the driven pulley spring tension. But I think the effect would be minimal and the application would be complicated.
And once a CVT is properly tuned to the power characteristics of the motor/engine, there would be nothing to be gained by altering it.

BTW, the CVT is an American invention, developed by an engineer working for the Salsibury Motor Scooter Co. in the late '40's. In fact, when I was first introduced to the system on snowmobiles in the '70's, it was reffered to as a Salsibury transmission.
Honda was the first to bring it into the realm of wheeled vehicules with their 4-wheel off-roader Odyssey in the late '70's. The term, CVT was coined by Honda with the U.S. introduction of their motor scooters in 1983.

EDIT-One way to control the ratio manually would be to "fix" the drive pulley and use a capstan/idler pulley somewhere along the belt routing to regulate where the belt rides in the driven pulley. My dad built me a Briggs and Stratton powered tractor in the 50's that used that set-up

[gwhy!]

I know the burgman 650 scooter has a manual override for its CVT and its done with stepper motors ( I think it has 4 modes , fully auto,sports,touring, and fully manual ) and the only true way of riding it for best Fuel eco and performance is in fully auto mode.

[Harold in CR]

I had a Crapsman Lawn tractor once. It had a variable speed for climbing-descending. Didn't change speeds a LOT, but, it was a cheap tractor. There was NO hydraulics that I know of. Maybe search for exploded view of a couple lawn tractors and see how the CVT is actuated ??

Had a JD 95R Combine with the CVT. It was actuated by a long skinny hydraulic cylinder. Might search for something like that, to see how it is rigged up.

[Solcar]

I've seen one type that uses two cones and a belt or chain between them and as the belt is pushed left or right, it goes toward the skinnier end on one and the fatter end on the other. The pulley method that Motomech described uses v-belts under a similar general principle, but with narrowing one pulley while widening the other. There is another way that I am also considering that might use a polyurethane cone and an abrasive wheel pressed against it. The abrasive drive wheel would be moved in for gearing up and toward the outside for gearing down. Right now, the third way is the most attractive option to me, except for the potential wear of the polyurethane cone. So I'm not planning any longer to post a side view of that first diagram, but here is a simplified diagram of the third way.

I wasn't able to find out something about that pulley on the Craftsman, yet I would guess it did it by using a centrifugal CVT actuator on the crankshaft pulley, with the idler pulley taking up the slack when the crankshaft pulley is wider at lower speed.

I have thought about that set-up Motomech mentioned that his dad did, and though I'm unsure how it really worked, it seems pretty ingenious. In those days, people were pretty good at coming up with innovative gadgets.

[Harold in CR]

The one I was referencing had a lever on the dash board. Pull it toward you and it would slow. Push it away and it would speed up. It was manually operated, somehow. I have a couple ideas, but, need to draw them up, to see if I'm thinking correctly.

I'm thinking, typically, the driven pulley is spring loaded, so, as the drive pulley expands from RPM's, it forces the driven pulley to move closer together.

If so, why couldn't it be rigged, so, you manually opened-closed the driven pulley, and the drive pulley would work as the driven USED to ??

It takes away from the automatic adjuster, but, it would allow the operator to do the adjusting. Emotor RPM's are much lower than 2 smoker gas engines.

I am planning to experiment with this idea, IF I ever get started on my 2 passenger trike. I have an 8000 RPM motor rated at 20HP at 30V. I plan to run it 72V, maybe. I would like to be able to gain regen to save brakes and get a little boost to the batteries, if possible.

[johnrobholmes]

An easy way to change the ratio of a CVT is to jam a thinned rollerblade wheel into the front pulley to push the belt lower. Super easy to set up on a pivot with lever actuation. The CVT will work how you tune it otherwise. Race scooters used to do this until it was banned for being too much advantage.

[Craigh85]

Something around this size might do for an ebike, think this was homemade for an r/c car. Pretty clever if nothing else.

https://www.youtube.com/watch?v=W0BERPIp364

[motomech]

The Honda Spree[and Yamaha Razz] were the smallest and lightest motor scooters ever offered,

honda-spree-05.jpg (36.24 KiB) Viewed 225 times

But they didn't use a true CVT, instead of a frt. variator, they just had a fixed pulley

SpreeAftermarketGear9.jpg (72.27 KiB) Viewed 225 times

SpreeAftermarketGear9a.jpg (68.56 KiB) Viewed 225 times

But the driven pulley was spring-loaded and moveable, to tension the belt. I suspose a guy could use a skateboard wheel in the manor that Johnrob described, to push down along the top run of the belt to draw the belt further down into the driven pulley. Not sure how far the driven pulleys separate, but of course, the range from low to high would be limited. At best, half of a true CVT.

The Honda Express ll was a No-ped that used a fully functional CVT and it'd belt was longer perhaps making it more suitable to an Ebike Application.

expess ll.jpg (61.35 KiB) Viewed 225 times

It felt quite different than the CVT's on scooters, which are seemless in opperation. It felt very much a 2-speed, bumping into high gear, not unlike a car auto with a shift kit. As I recall, the Express ll[and simular varients] used steel for all the components, making the system heavier than the ones on scooters.

Bear in mind, that all these scooter and No-peds use a reduction gear final drive in the rear part of the housing.

[ddk]

my old Honda PA50 used a cvt variable-width pully system
worked ok for a few thousand miles then gave me fits until I finally sold it.
All cvt designs, even the nuvinci, suffer from frictional loss, making them perhaps not so great as a final drive for something human-pedal-operated.
I've thought of implementing something like a dual cage derailleur like the pic.
Should be easy to implement on a recumbent with those stupidly long chain lines.

9000 hours in inkscape

dualing deraileurs.jpg (5.26 KiB) Viewed 223 times

suffers one more friction point (jackshaft)

pedal bike with 36 gears choices.jpg (5.54 KiB) Viewed 221 times

[Solcar]

There are a lot of ideas with potential, and reading about them for the last few posts brought back some memories. I had a "Yamahopper" what seems like a long time ago and it indeed was small in size.

The image of the two oppositely-oriented rear clusters with a shiftable chain between them reminded me of our friend, Numberonebikeslover, who was planning to try to regulate his speed by only the changing of gears rather than by electronics, if memory serves me. All that adversity between groups where he lives keeps things pretty awful, it seems, and I recall the electric power being off for large periods of time.