Core77

ID student Chris Holloway's invention one-ups bicycle engineers

By Rain Noe - July 16, 2010

in Tech | Object Culture

We so often see design done by engineers, and rarely see engineering done by designers. But Chris Holloway, an industrial design student at London's Brunel University, has pulled off a rather impressive feat of ID's sibling-rival profession.

"While every other part of the modern mountain bike has changed to meet the modern demands of downhill racing, the gear system has barely changed for more than 100 years. For the pinnacle of the sport, it is clear that a better alternative had to be found," said Holloway, providing background for his invention: The Link gear system, which rather than being a series of fixed-size gears is instead made of four moveable segments.

By shifting inward or outward, the segments essentially change the diameter of the gear and thus the gear ratio. Beyond that I have absolutely no frickin' idea how it works or how he could have possibly cooked up something this complicated-looking. And Holloway even designed it so it could be retrofitted to existing cranksets, so you don't have to get a whole new bike to enjoy his system.

"On full suspension bikes, it is important to minimise unsprung weight to obtain the maximum performance. By removing the rear derailleur and cassette, the mass is relocated between the riders feet and away from the suspension, which gives much better small bump response, as well as increased grip and cornering ability," Holloway explains.

Is the idea both workable and mass-produceable? It sure has Brunel University's vote of confidence--they've patented the design.

via the engineer

o

Favorite This
Q
24

Comment

Comments

Emily

11 years ago

Z

Z

Reply

I have tried the gear assembly in Solidworks, the gears still engage to each other with missing a few tooths the gear still works.

We do not need to make full teeth gear.

!Report as spam
Laz

12 years ago

Z

Z

Reply

There was a Polish engineer that came up with this idea in the late 80's to early 90's in the UK and he wasn't the first to try it. His idea used only one axis though, giving an expanding Ovaltech/Biopace type of ring. The weight of it only countered the weight of the derailleurs. So nothing was gained, except a huge increase in price.
Best wishes and good luck to Chris Holloway. I'd love to see something work that is a bit less "brutal" than throwing a chain off its ring. Just remember half of your market will experience snow, grit and/or mud. ;)
!Report as spam
C L

14 years ago

Z

Z

Reply

Actually, it might work, if the segments tilted for a split second during each shift.

Also, I've noticed a few bikes out there with belt drives. It seems this system might work with a belt drive arrangement.
!Report as spam
Dave

14 years ago

Z

Z

Reply

This design has been invented, patented, reinvented, and rehashed hundreds of times. Take a look back into the history books, cool thinking, but nothing at all new here.
!Report as spam
andyc

14 years ago

Z

Z

Reply

Fascinating. I could see that the problem with the chain not stretching/compressing could be overcome if each section could move in series, rather than all at once. So the first unengaged section (the section of the ring above the rear roller) could move in or out; you keep on pedalling; that section would hook up with the chain and the next unengaged section would then move in or out as well, and so on until all 4 sections have moved. Not sure how that would feel when riding though- but possibly not all that different to current set-ups.

But yeah, to unnecessarily point out the apparent problems with this product: it does look like mud would kill it, and that spring at the front looks awfully fragile. Also, is that thing at the top supposed to be a chainguide (the bit with 'Link' written on it)? If so, it should be on the other side of the 'chainring', shouldn't it? Also looks to have fairly limited range of gear ratios for DH too; could see it maybe working better on a commuter bike or something, but then the thing wouldn't be any help to the suspension ('cos there wouldn't be any) and you may as well use a Rohlhoff hub...
!Report as spam
Jones

14 years ago

Z

Z

Reply

@Ed R
I was also going to mention the Rohloff hubs. This article makes it sound like drivetrain designs have always been monopolized by the dual derailleur design. The epicyclic hubs date from the mid-1880s.
http://en.wikipedia.org/wiki/Rohloff_Speedhub
!Report as spam
Nerobro

14 years ago

Z

Z

Reply

for all the complaining about first year engineering students, nobody has thought that his might actually be a good design . No, chains don't stretch, but there's 1/2 of the crank rotation where there's no chain on the sprocket. If he uses a racheting/spring loaded sizing system the "size" adjustment could be input, and the chainring would reset it's size during the next full pedal stroke. As each segment left the chain, it would snap out to the chosen size.

there are other ways of doing this too.
!Report as spam
Greg

14 years ago

Z

Z

Reply

Perhaps an answer to the "stretchable chain" issue would be some sort of cam being used for the gears to change shape? Who says the entire circumference of the gear has to change all at once? The sprocket could start to expand at the point where the chain first engages the assembly. This expanded segment would then rotate around the gear, with the chain behind it engaging an equally expanded segment of the ring. (remember in a standard shifting setup that when you shift to a larger gear the chain doesn't automatically jump onto the larger sprocket, but rather grips a tooth and then fully engages as you peddle).

I'm an engineer and while I love that people are trying to improve existing designs, I just don't think this will be practical in real life. Industrial designers come up with great concepts, and really nice looking renderings but often don't consider the real world applications. What will happen when this thing gets wet? Throw a little mud and grit into the thing and take it out for a 20 mile ride....still work? What happens when it breaks out on a ride, can I bypass it and still make it home like I could with a regular derailleur? Maybe they've considered this and accounted for it already. As people have stated this design is still early...lets hope they can work it out!
!Report as spam
Ben

14 years ago

Z

Z

Reply

I have also thought about this concept a lot albeit in the context of making racing bikes more aerodynamic (shoulda woulda coulda). I believe there is an obvious answer to the issue of not having a stretchy chain. Rather than have the four sections of chainring expand at once they should move outwards (or inwards) one after the other as they pass the "9 o'clock" position. Each section will be unengaged from the chain at that point and free to move. Upshifts could also therefore also be made with less force as they would be free from chain tension.

For all I know an idea like this is built in to the concept but either way I find some of the other comments on this thread (I'm talking to you bemused engineer) extremely condescending.
!Report as spam
Chris Holloway

14 years ago

Z

Z

Reply

Hi everyone.

Thanks very much for all of your comments on the gear system! Most of you have got the idea of the concept, although it is difficult to see how it works with just this one photo (for a few more, have a look on http://www.messdesign.co.uk/).

As many of you have pointed out, chain is not known for its stretchiness, and this was the main challenge with the project. Each segment expands independently in the space where the chain is not in contact, so that when it meshes with the chain again it is at a larger diameter, therefore giving you a larger effective gear size.

The idea of this prototype is to show that eventually, you could develop maybe a 7 or 8 speed version which would hopefully eradicate the need for any sort of rear derailleur. This is impossible to achieve with planetary systems like the Truvativ Hammerchmidt, which are fundamentally limited to only 2 speeds. Geared rear hubs unfortunately aren't an ideal solution either, as their huge rotating weight causes a gyroscopic effect in the rear wheel and has a large, unwanted effect on the suspension system. Every system has it's problems, so this, for me, is a compromise of a number of factors that allow it to work as well as possible with any modern downhill mountain bike.

The prototype does prove that the 'principle' works, but there is an awful lot of development work and testing still to do in the future.

Once again, thanks very much for all of your constructive comments!

Cheers,

Chris Holloway
!Report as spam
Ed R

14 years ago

Z

Z

Reply

While it looks interesting, and represents an unconventional approach, maintaining chain tension is going to be a major headache and mud is going to destroy all those moving parts. There are some very good reasons why variable drivetrains like nuvinci focus on on the rear hub where diameter changes are smaller and why mountain bike drivetrain development is focused on enclosed planetary gear sets like Hammerschmidt cranks and Rohloff hubs.
!Report as spam
bemused engineer

14 years ago

Z

Z

Reply

What sensationalism! This design hasn't "one-upped" anyone until it works, which it almost certainly doesn't. Really, when are designers going to leave the poor bicycle alone?

Previous commenters hit the reason for its non-functionality on the head: as the segments shift outward there's no way the chain can stretch to accommodate the change without some kind of horrible chain skip or actually being lifted from the chainring (so you can forget about shifting under power).

This reminds me of the kind of project put forth by freshman engineering students, full of ideas but lacking the discipline to fully examine the shortcomings of their design: the kind of thing that professors sigh at and hand out B-minuses to for effort in spite of naivety.
!Report as spam
slippyfish

14 years ago

Z

Z

Reply

SRAM is doing something similar with their all-mountain or downhill-specific HammerSchmidt, a "single-ring, planetary-drive crankset", which can go from a 22t to a 36t equivalent without the chain-stretching issues that this concept faces. Matt, I hope that Chris keeps working on this concept too, but we don't need a lecture about "remembering he's a student, no solution is perfect," etc. because right now the concept doesn't work.
!Report as spam
Matt

14 years ago

Z

Z

Reply

I'm a ex-brunel graduate and can see the beauty of the concept. I visited the New Designers show and looked at the prototype in real life. It is a labour of love and thing of engineering beauty. A lot of hours have been spent in the workshops creating a long line of prototypes. The mechanism reminds me of the old YouTube videos for the expanding table. http://www.youtube.com/watch?v=Bh_qn62zny0

There are a number of limitations and issues with the concept, that have been brought up a number of times by different people in their comments, however everyone must remember that Chris is a final year student who has a limited time period (approx 6 months) to create his final year project. This is the first iteration of a concept that has overcome a huge number of problems, but has served to identify plenty more. When approaching problems from a design perspective - no solution is 100% perfect but what you learn as you develop will in the end mean that the final solution is "perfect enough".

I hope Chris has the opportunity to take this on further and actually be involved with developing a production version that may be 4 or 5 iterations down the line.
!Report as spam
Shaw

14 years ago

Z

Z

Reply

It seems like a good idea, although I'm not sure how an increase in diameter is accomplished with the teeth engaged in a chain, the circumference distance seems solidly fixed. The chain seems like it would have to be lifted out of the sprocket before adjustment was made to the diameter.

It also seems to have a lot of precedents in the patent filings. 4521207 for example 1985, 3995508 from 1976, and the others referenced.
!Report as spam
James Willson

14 years ago

Z

Z

Reply

I'm at brunet and it's awesome to see great designs like this getting the exposure they deserve.

James
!Report as spam
Lee Gibson

14 years ago

Z

Z

Reply

I'd be interested to see how they account for the fact that chains aren't stretchy, so it has to somehow allow for the segments to be moved one at a time, and then adjust after all four segments are set to the second diameter, and then not feel like you're pedaling in squares (unless we want to go back to the BioPace days).
!Report as spam
Benjamin Guerrero

14 years ago

Z

Z

Reply

I think i understand how this system works, and i see one issue about it, when you are on the bike moving and the gear is rotating, you will need an elastic chain for the system to work or it will be freezed or the chain will brake
!Report as spam
Jim N.

14 years ago

Z

Z

Reply

As the toothed segments contract towards each other, you'll get a Biopace-like effect where the chainring becomes rounder.

Never having downhilled, I can't say if that would be a problem or an advantage.

It looks like with each repositioning of the teeth, you get a 4-tooth jump in front gear, with a minimum of 40. So with the ability to move 3 times, your effective (biopace) chainrings are 40, 44, and 48.
!Report as spam
Bryan @ Renaissance Bicycles

14 years ago

Z

Z

Reply

I think the concept is great, but I am skeptical that it will work.

I created a similar prototype a few years ago ... but the flaw in my design was that there was no way for the chain to "stretch" as the chainrings expanded. In other words, with the chain meshed to the teeth, it couldn't expand with the outward movement of the chainrings. The chain would grip on the teeth and then eventually "skip" to work with the expanding diameter.

I see there is a jockey wheel / tensioner, but this solution didn't work for me either. Hopefully the designer has a clever way to deal with this issue.

!Report as spam
Al

14 years ago

Z

Z

Reply

Breaking it up into segments of fixed spacing and segments of empty space is a great insight. I assume it isn't a CVT, but has fixed setting notches. The length of the gaps has to be held to an integer multiple of omitted sprocket teeth so the chain will engage.

There is one problem I can see. The loop of chain over the sprocket is roughly conical, so slightly over half of the teeth on the front sprocket are engaged and sharing the load at any time. In this design, the same teeth are bearing the load at all gear ratios, so at higher gear ratios more force is being applied to the same number of teeth and the same number of chain links. Might lead to increased chance of broken chains under high load. If the bike is really made just for downhill, then no problem, but if you want to climb a hill with it...
!Report as spam
Ron

14 years ago

Z

Z

Reply

"On full suspension bikes, it is important to minimise unsprung weight to obtain the maximum performance. By removing the rear derailleur and cassette, the mass is relocated between the riders feet and away from the suspension, which gives much better small bump response, as well as increased grip and cornering ability," Holloway explains."

By how much. Lets have that.
!Report as spam
Salvador Morel

14 years ago

Z

Z

Reply

Just magnificent Industrial Design thinking; I must say when I read the title of the article I had my doubts... but thru the concept I really understood how much importance re-thinking basic products structure has, here we have somebody who knows how to do it right... CUDOS to the designer!!!!
!Report as spam
Lyle

14 years ago

Z

Z

Reply

While I would gladly cheer getting rid of both deraillers, the existing gear system hasn't changed in years because it's reliable and simple. The fewer moving parts, the better, and this is one huge moving part. How would it stand to being bashed against a tree? Eager to hear more. Interesting, but skeptical ;)
!Report as spam