Note de ce sujet :
  • Moyenne : 0 (0 vote(s))
  • 1
  • 2
  • 3
  • 4
  • 5
A true planetary CVT
The following post is part of a topic copied from the discontinued NuVinci Forum. The majority is written by members of the old forum and not by me.

I found this discussion interesting and thought I would post it for anyone who is interested. I have removed the first part as it isn’t the most interesting bit and CadenceKing summarizes it well in his post below.

by CadenceKing on Fri Oct 19, 2012 4:25 pm

As this thread is approaching its one year anniversary, I have just read through all of the comments and studied all of the videos referenced. Here’s my evaluation of what has transpired.

Myrtonos began this thread with the provocative proposal, “. . . how about continuously variable hub gearing?” Referencing planetary gearing systems, Myrtonos points out that in existing gear systems, “different ratios are given simply by locking and connecting different components.” It is at his next point that the engineering difficulty creeps in. Myrtonos states, “But the gear ratio of a planetary gear set can be varied simply by varying the amount of slip between two of the components.” (Note: This is not standard usage of the term ‘slip’ and would be strongly objected to in engineering circles. Myrtonos later explains that he uses the term ‘slip’ to refer to “speed difference between two components.”) Here’s the real problem: no concrete details are given regarding the ‘invisible hand’ (my term) that would cause the speed differences between components!

Bigoilbob has repeatedly pointed out the inevitable power losses associated with the ‘invisible hand’s' intervening with the relative speeds of the gear train components. (I’m confident that Bigoilbob would agree with me that if one had a planetary gear set and started changing the speeds of certain components (i.e., the sun or planet carrier) that a variable input/output ratio would indeed result. . . ignoring efficiency issues.) He eloquently argues that the big problem with the concept is the loss of energy/efficiency!!!

Until such Ill-defined expressions as, “hub gearing to subtract the speeds of the input and output discs” and “ratio variator” are documented with specific component geometries and hardware, the “true planetary CVT” remains in the same realm as the ever elusive perpetual motion machine.

I empathize with Myronos that it can be fun to visualize a world where gear systems mesh with each other in interesting loops and paths! But at some point hand waving and words must give way to real sketches, drawings, working models, prototypes, etc. An “invisible hand” slowing or speeding up components won’t pass muster at an engineering concept review. (Nor will nitpicking over terminology.)

Don Miller took a clever variation on the theme of the planetary gear train and shepherded it through the above developent stages all the way to production. The resulting hubs function in the real world (I just got back from a ride on my NuVinci-equipped Dahon Cadenza!) at a very reasonable level of efficiency.

I hope this helps!

Cadence King

He's a smooth operator,
smooth operator,
smooth operator,
smooth operator. . .

Re: A true planetary CVT
by Myrtonos on Fri Oct 19, 2012 6:47 pm

bigoilbob wrote:I looked at the video, and maybe I'm missing something. Where is the continuous variability and how is it controlled, without some kind of energy sapping slip mechanism?

It's controlled in the same way as with the NuVinci Hub but the plantetary hub gear subtracts the ratio of two of the components. One idea is for the input disc to drive the sun gear, it's teeth acutally surrounding the disc and the planet pinions attached to the ouput disc, the ring gear being attatched to the bicycle wheel. The idea is to use a NuVinci type ratio variator to vary the speed difference between the sun gear and the planet carrier.

Re: A true planetary CVT
by bigoilbob on Fri Oct 19, 2012 8:58 pm

Right on Cadence King, except for that "eloquence" part. I would like for someone who has taken their engineering statics and dynamics courses less than 35 years ago to provide a better explanation/quantification of the lost power than I did.

Re: A true planetary CVT
by Myrtonos on Wed Dec 12, 2012 9:21 am

This page will help, it clearly shows the a belf drive as a ratio variator, and a plenetary gearset that subtracts the speeds of the input and output cones.

Re: A true planetary CVT
by bigoilbob on Wed Dec 12, 2012 3:37 pm

Hi Myrtonos;

As B4, I agree that you can change the ratio of a planetary gear set infinitely, by allowing other parts of it to move. But what we disagree on, and what I am still failing to see, is how you can do so without, AT LEAST proportionate power loss.

Let me try another explanation. Power, as energy, has to be conserved while transmitting it. Some can be lost to another form of energy transfer/unit time (such as heat produced at a certain rate), but it must be conserved. So, if you reduce the speed of an output shaft in a variator, say by half, that torque will not increase, but the speed will drop by half. So, half the speed, same torque, half the power. Where did the rest of the power go? It went into the device you used to keep the other part of it (the part you are allowing to move) moving at the speed you chose - to drop the speed of the output shaft by half. If that is a brake, it is lost as heat. If that is a generator, it is "lost" as electric power. The "lost" heat rate or electric power generation rate might be harnessed elsewhere, but energy in/unit time=energy out/unit time. Always. In this case, half of the power is kept as mechanical power, half is "lost". Bad news for motorized bike builders like me.

But if I am missing something, please point me to the energy flow into and out of the device that I am missing. I genuinely enjoy your posts, and admire the thought you put into them.


Re: A true planetary CVT
by Myrtonos on Wed Dec 12, 2012 4:46 pm

The rest of the power does not go into that ratio variator, but through it. And in that diagram, it is not a brake but a CVT belt drive.

Re: A true planetary CVT
by bigoilbob on Wed Dec 12, 2012 6:36 pm

Thanks Myorntos;

I'm with you and I also did not look at your drawings close enough. I need to work on that.

But that being said, isn't this just a variable belt drive hooked up to a planetary drive, with the inherent inefficiencies and Rube Goldberg construction? That is, not the "true, infinitely variable, planetary drive" that would have indeed been an eye opener. I checked out all of your links and could not find an example of a real working unit or model. I might have missed it, so maybe you can aid my search. For infinitely variable, reversible, drives we have been using variable swash plates for a long time. When I was an equipment operator in the Seabees many, many, years ago, my vibratory roller had one. Would your drive be as good or better than those?

Re: A true planetary CVT
by Oran on Sun Dec 16, 2012 6:26 pm

I really like the idea of combining a CVT with a planetary gear set to create an IVT I think it might have potential. It wouldn’t be much use on an ordinary bicycle but on a cargo tricycle or pedicab it could be really good. The incredibly low ratios that are possible would be ideal for getting a heavy load moving and reverse would be a bit of a novelty.

I might have a think about this over the Christmas holiday, calculate some ratios, work out how the power splits and decide which of the 6 options will work and which is best.
There are lots of questions to be answered.
How much power will be transferred through the CVT? At what rpm and torque?
Will the ratio range of a NuVinci hub be too low or too high? If so can this be compensated for by using different ratios in the planetary gear set?
Could the planetary gear set replace the freewheel? Can the CVT remain in the wheel? I can’t imagine how this will work when in reverse.

At the moment I’m thinking that the power transferred through the CVT may be mostly as torque in which case a NuVinci hub will be inefficient. Whether this is an issue will depend on how much power we’re torching about.

I can’t help but think that there must be some reason why this won’t work.

Re: A true planetary CVT
by Myrtonos on Sun Dec 16, 2012 6:41 pm

Bicycles don't ususally need to back-up, so one could make geared neutrual the lowest ratio.

Oran wrote:How much power will be transferred through the CVT? At what rpm and torque?

The portion would vary accroding to the gear ratio.

Oran wrote:Will the ratio range of a NuVinci hub be too low or too high? If so can this be compensated for by using different ratios in the planetary gear set?
Could the planetary gear set replace the freewheel? Can the CVT remain in the wheel? I can’t imagine how this will work when in reverse.

One could make the ratio range of the variator itself narrower, becasue the planetary gear set expands these ratios, and becasue the variator deals with less torque. If the IVT were to replace the freewheel, the rider wolud have to shift very rapidly into neutual, and the pedals would turn unusually fast during that moment.

CVT + planetary gear set = IVT
by Oran on Mon Dec 17, 2012 5:22 pm

Myrtonos thanks for your reply.

I thought of some more questions after I submitted the previous post. You don’t have to answer all these questions I’m just putting my thoughts into words.

Something I was wondering was does the CVT need to transmit torque in both directions as a NuVinci hub can’t? Am I right in thinking that if reverse isn’t used then it shouldn’t have to? I suppose it is also possible to avoid going all the way to neutral and so avoid the pedals turning too fast.

The other thing I am trying to work out is what’s happening at very low ratios. Which of the two routes CVT or planetary gear set will be taking the majority of the torque? For this idea to have any chance of working the NuVinci hub must be kept within its torque limits. Currently this limit is maintained by having a minimum sprocket ratio. Would an IVT be a way of getting lower ratios without over stressing the CVT, if so this could increase the rather poor efficiency currently experienced with a NuVinci hub in the lowest ratios.

You mentioned that the planetary gear set expands the ratios so I assume this means that you could get more than the current 360% without having multiple chainrings. Multiple chainrings have the disadvantages of needing a chain tensioner and ruling out the possibility of a belt drive.

What I’m thinking is that this idea could help a NuVinci hub perform better in areas where it currently falls short. Applications such as cargo tricycles, pedicabs and possibly mountain bikes. The sort of applications where a CVT or IVT could have the biggest advantage.

Re: A true planetary CVT
by bigoilbob on Mon Dec 17, 2012 7:41 pm

A NV hub could theoretically be harnessed to the planetary drive to put the P drive into a forward ratio, neutral and reverse, the same as a belt driven CVT. Maybe even better because the NV's can ratio 3.6*, v the ~2* for most belt driven CVT's. I can't internalize the NV torques any more then the belt driven CVT torques - you would have to break down and do free body diagrams for that. But I think NV puts an r/m limit of 1000/minute and if they are used as bike hubs they probably don't get spun much more than 1/3 of that. So, if you designed your NV hub/planetary gear device properly, you could transmit over 3* the power you could thru the NV hub, than if it was used as a bike hub.

Fun to write about, but just musings. A CVT/planetary gear combo has no practical use that can not be better met from current assemblies, as far as I can tell.

Re: A true planetary CVT
by Oran on Tue Jan 08, 2013 4:42 pm

I had a look at this report by Andrew J. Fox I found on the internet and it appears bigoilbob might be right.

It seems that it is not possible to increase the ratio range of the CVT and reduce torque through the CVT at the same time. This probably explains why this arrangement of CVT and planetary gear set is not widely used.

In theory it is possible to reduce the torque through the CVT, enabling lower ratios and increased efficiency but the ratio range will be much reduced. Multiple chainrings would be required or the top speed will be very low. I don’t think this will be any better than a mid drive arrangement.
Using the planetary gear set to increase the ratio range of the CVT probably isn’t an option as it increases torque on the CVT. This is due to some of the power recirculating within the transmission.

A CVPST (continuously variable power split transmission) or IVT is an interesting idea and I’m sure somebody could use it in some unusual situation but in reality it has limited potential.
- Oran
I’ve just come across Fallbrook’s latest video NV TechVideo Final 1280x720 and it was worth watching through all the stuff I already know because at 2:20 it got more interesting. It explains how everything that has been discussed above can be achieved within the transmission and without the need for additional planetary gear sets. I never thought of that.

I suggest you skip to 2:20 unless you have no idea what a NuVinci hub is.
- Oran
Great video! I wonder what kind of power lost are involved in these alternate configuration...
Having the balls spinning the rings in different direction while producing torque will certainly
result in high friction and head... clamping pressure will have to be immense...

But when you think about a gasoline engine, those power loss are negligible when compared
to what you gain by running the engine at optimal RPM/torque/efficiency at all time!
Si ça a déjà été fait, je peux le faire
Si ça n'a jamais été fait, donnez-moi juste le temps de trouver comment !


Atteindre :

Utilisateur(s) parcourant ce sujet : 1 visiteur(s)