Roller Pinion + Shaft Drive

[Mr Evil]

In a change from overpriced bearings, CeramicSpeed have shown a drivetrain consisting of a shaft driven by a roller pinion, which they claim can reach 99% efficiency. Compared to the 98% of a typical chain drive, that might not sound like a big improvement, but that's half the losses, which is actually quite impressive.

Although it looks impractical now, I think the most interesting thing about this is that it might renew interest in shaft drive, which has never caught on before, partly due to low efficiency. I've never seen roller pinions used before, and there isn't even a Wikipedia page for them, despite having been invented a long time ago. There is some information about real-world use though, which corroborates the efficiency claims.

[Brucey]

very interesting. I got quite excited when I saw this photo


until I read in the text that this is just a mock-up that, ahem, doesn't actually work. All the tests have been carried out on a singlespeed transmission to date.

It is quite a clever idea but its claim to high efficiency may yet prove to be an elusive dream. They don't say so, but my guess is that they (like most others) have used a constant torque to measure efficiency to date. This is probably the lesser of several evils (potential sources of error), provided the transmission is stiff enough (which with a chain is easy enough to achieve). The problem is that this transmission almost certainly isn't going to be stiff enough. There are two main problem areas as I see it;

1) the shaft itself may be nowhere near torsionally stiff enough (shaft drives on bicycles usually aren't) and

2) the thrusts at the 'chainring' and the 'sprocket' are nowhere near perfectly tangential; they contain an appreciable axial thrust component.

This image shows that the axial thrust is likely to be larger than the transmitted tangential thrust in some cases


The axial thrust will stress everything in a direction in which it is not (and for various reasons cannot be) made to be terribly stiff. As soon as any kind of real-world pulsy torque is applied, I think the parts will flex horribly in a way that won't conserve energy very well.

I wish them well with it, but I think this is another system that won't actually pass muster in the real world.

cheers

[hamster]

I loved the bit about how they haven't actually worked out how to get it to change gear yet.

Figuring how exactly how that rear pinion would move falls outside of CeramicSpeed’s area of expertise, though

[SA_SA_SA]

I saw a similar single speed shaft drive idea in A to B magazine a fair while ago .

Surely a more efficient 'single speed' shaft drive is still of interest to some hub gear users?

[Cyril Haearn]

If it were as robust as a chain drive and maintenance-free I would like to try it
Maybe it will work in the end and replace chain drives altogether
There are a few bikes with shaft drive already in use

[Brucey]

I saw a similar single speed shaft drive idea in A to B magazine a fair while ago .

Surely a more efficient 'single speed' shaft drive is still of interest to some hub gear users?

why? Losses in a chain are often a small fraction of those inside the hub, so efficiency is not a good reason to do things differently here. It is very easy to protect the chain and give it a long service life by putting in a chaincase, so longevity is not really an issue either.

These points explain nicely why the (heavier, less efficient, often less reliable, always more expensive) shaft drive has never been popular despite several goes over the last 120 years.

The system just announced might appeal to derailleur users if they make the gears work and if the efficency in reality is in any way comparable with a chain in real life (rather than under constant torque).

cheers

[SA_SA_SA]

.....Surely a more efficient 'single speed' shaft drive is still of interest to some hub gear users?

why? Losses in a chain are often a small fraction of those inside the hub, so efficiency is not a good reason to do things differently here. It is very easy to protect the chain and give it a long service life by putting in a chaincase, so longevity is not really an issue either.

It could perhaps avoid long chain runs on recumbents.
You can't put a chain case on a Brompton, but perhaps a detachable/extendable shaft?

For better or worse manufacturers might fit a shaft drive whereas they seem happy to fit bare hub gear chains.

[Vorpal]

I agree with Brucey, at least with regard to the load on the roller pinion. In order to maintain the high efficiency, it also requires considerable precision, which I think will be difficult to achieve, between the drive shaft stiffness (it is out on a lever arm that is long relative to the load required) and the dirt and grit that will be introduced in use.

It may be possible to overcome these problems, but it is, IMO, nowhere near a practical solution.

[Mick F]

The bicycle as a basic design of which we have now, was invented a century ago.
They still haven't improved that design despite many many ideas.

This is just yet another dead end IMHO.

[fastpedaller]

Unless the drivetrain is protected from the elements any system will suffer from similar issues as the chain drive. A gearing system in the bottom bracket or hub (as has been done many times before) is nicely enclosed, and coupled with a more robust chain is maybe as close as possible? A chain cover/oilbath completes the job?

[hamster]

I agree, just imagine the effect of muck on all those teeny bearings...

[alexnharvey]

Unless the drivetrain is protected from the elements any system will suffer from similar issues as the chain drive. A gearing system in the bottom bracket or hub (as has been done many times before) is nicely enclosed, and coupled with a more robust chain is maybe as close as possible? A chain cover/oilbath completes the job?

I am sad that integrated chainstay-cases have not taken off more quickly

[Vorpal]

Unless the drivetrain is protected from the elements any system will suffer from similar issues as the chain drive. A gearing system in the bottom bracket or hub (as has been done many times before) is nicely enclosed, and coupled with a more robust chain is maybe as close as possible? A chain cover/oilbath completes the job?

It would suffer from worse issues than a chain drive. A chain is not especially sensitive to road gunge. It just wears faster. A chain drive system may lose a little eficiency, but it has to be *really* bad before it is noticeable to most users.

The roller pinion system will be much more sensitve to road gunge, and unless they make it a closed system, IMO, it will take very little use on the roads before the efficiency is reduced to noticeably worse than a chain drive, even if they can overcome the other issues. They could overcome this problem with a much higher force at the interface, but that would require larger parts and/or additional stiffness in the drive shaft, which soon becomes impractical from a weight perspective.

[Norman H]

YouTube video here.

[Samuel D]

YouTube video here.

There’s visible backlash at 1:37 (more easily seen at 0.25x playback speed; link starts five seconds before the backlash).

Rack and pinion machines are often preloaded to avoid backlash, but doing that here would eliminate the possibility of shifting with the scheme they envisage.