Tuesday, October 09, 2007

Is it time to re-think the derailleur?

One area of the bicycle that seems to be overlooked when it comes to equipment performance is friction; especially in chain drives and derailleur gears.

From an engineering standpoint the derailleur gear is an inefficient system. However, since its invention in France in the 1920s, no one has come up with anything better.

The system of a chain drive running out of alignment on multiple sprockets is not used on any other application except a bicycle as far as I know.

When the chain is not pulling in a straight line, there is extra fiction on the side plates and the bearing pins of the chain. There is also friction on the sides of the teeth on the chainwheel and rear sprocket.

When a chain is in alignment there is only slight friction on the bearing pins as the chain goes around the top portion of the rear sprocket and chainwheel. There is little or no friction on the side plates of the chain.

One thing a person notices the first time they ride a single speed fixed gear bike, is the smoothness of the transmission and the lack of friction. This is because the chain is in alignment, and there are no pulleys the chain has to run around.

The pulleys on the rear derailleur are the other source of friction; there is the friction of the pulleys themselves, and the chain has to go around a constant "S" curve. Turning the links of the chain, first in one direction, the other.

One derailleur popular in the 1930s and 1940s was the Osgear. (Left.) It had chain tensioning arm with a single pulley just under and slightly behind the chainwheel.

This meant the chain ran in the same direction and was not made to go around an "S" curve; there was also one pulley instead of two. At that time freewheels only came in 3 and 4 speed.

The Osgear had its shortcomings; it would not work with a double chainwheel because the tension arm was fixed. However, had it pivoted on a simple ball joint and had sideways movement, it would have aligned itself as the chain switched from one ring to the next.

The other drawback was, the fork that shifted the chain on the rear sprockets was over simplistic and shifting was not that good. Had it been designed like a modern front derailleur it probably would have worked much better.

A modern front derailleur is very efficient in that it will shift the chain over a ten teeth span or more, and once it has shifted the chain it is no longer in contact with the shifter and so causes no friction.

The Osgear had fell out of favor by the 1950s when the French made Simplex and Huret derailleurs appeared; they shifted better, and worked with 5 speed freewheels and double chainwheels. The Simplex and Huret rear derailleur had the chain wrapped around two pulleys in the "S" fashion; the way all modern rear derailleurs are designed today. I do feel the Osgear was a very efficient design that was never fully developed.

Campagnolo’s Cambio Corsa derailleur (Below.) patented in the 1930s but developed in the 1940 was a masterpiece of engineering for its time, but extremely difficult to use. A long lever released the quick-release, the wheel moved forward on a rack built into the frame’s rear dropout thereby loosening the chain. Another lever shifted gear while back-pedaling. At the same time, the wheel moved back tightening the chain, and the quick-release was re-tightened. There were no pulleys to tension the chain, so no friction.

Was the Cambio Corsa developed to its full potential? Has anyone ever experimented with sprockets that slide sideways on the rear hub so the chain is always in alignment? It would not be necessary for the hub to be wider, or the rear wheel dished more; the hub could be large enough for the sprockets to slide inside. Another idea, fixed rear sprockets, and a chainwheel that freewheels, gears could be shifted while coasting.

If the chain is to remain out of alignment, how about a chain with spherical rollers at each joint so it will run out of alignment without the friction of the side plates. I know all these ideas will cost more, but with the price of the top of the line bike what it is today, what is a couple of hundred dollars more for a drive train with less friction that will allow a rider to go faster.

The derailleur gear has remained basically the same for over fifty years; all improvements have been in shifting and the number of gears. Friction is overlooked because you can’t see it; and if everyone is using the same design equipment it is not an issue.

In some of my recent posts I have waxed nostalgic and longed for simpler times. I am not against change if it benefits the bicycle and the cyclist. Many changes I see benefit the manufacturer, and then sold to the consumer after the fact.

I am just throwing out a few off the wall ideas that may or may not be practical, but would it hurt one of the manufacturers to put a little money into some research and development to find out just how much of the rider’s energy is wasted overcoming friction?


Anonymous said...


I've seen a bicycle with derailleur gears where the freewheel is in the chainset, allowing gear changing whilst coasting.

It's a Japanese childrens bicycle that belongs to my son (he's outgrown it), so it's in my garage at the moment.

If I move the tandem out of the way I can get it out and take some pictures if you like.

Chris H said...

It's unfortunate the industry feels there's more of a market for horrible ideas like electronic shifting versus efficient shifting and drivetrain performance. Maybe the only good thing to come out of the recent infatuation with the fixed gear bicycle is the desire from the consumer for more efficiency in their geared bikes.

Anonymous said...

Any thoughts, ideas, or ratings on the internal gears/rear hub systems?

Anonymous said...

"Working on a new, novel integrated Rear Hub and Control design employing 21st Century manufacturing and materials advances."
NOT a NuVinchi, Rohloff, Shimano or Sun Race (Sturmey-Archer) variant.
Learn from others BEFORE product creation!

jim g said...

In the late 70s or early 80s, Shimano had a "Front Freewheeling System" which incorporated a freewheel into the chainrings. The presumed advantage was that you could shift w/o pedaling. I've seen this setup on cheap Huffy, Free Spirit, and Ross bicycles from that era -- it was commonly combined with Shimano's "Positron" indexed shifting.

Ah, Sheldon to the rescue!

1982 Shimano Catalog page on FFS

1982 Shimano Catalog page on Positron

Marrock said...

What about friction issues with internally geared hubs?

I've always been curious about them.

Moveitfred said...

Cool discussion. I can't add anything to what's been said, but I can say that Vervecken was none too happy with his derailleur this past Sunday.

Anonymous said...

After reading the blog, I was going to ask what Dave brought up in his last paragraph: how does the friction affect the rider and has anyone studied that? Is the effect of friction overriden by the fact that the frames and components have gotten so much lighter over the years?

Finally, when I die, I'm going to ask God what the friggin difference is between affect and effect. It's no wonder I was a math major.

Anonymous said...

I have seen a guy riding a shimano positron in Vancouver BC. It was louder than heck when it freewheeled. Like a fixed gear, it's really dangerous if you don't roll your pants up out of the way of the gear, and it can amputate a finger when it's up on your workstand. I bet this had something to do with why this system was not more widely adopted.

Anonymous said...

Whoops, that's FFS, not positron!

VintageSpin said...

And how much more friction was created by the 10-speed system? Something engineers didn’t consider on their way to squeezing more sprockets into a 130mm hub spacing.
That friction also wears components out faster, chains and sprockets; especially at extreme chain line gears like 53 x 20 or 39 x 13.
Seems engineers/designers should stop the marketing ploys and make something really great. But then, what incentive do they have when they have gotten away with it for so long?

Anonymous said...

what about a cassette mounted on a gimbal that changes its axial drection as the chainline changes

VeloJon said...

Having recently re-read Frank Berto's great book "The Dancing Chain", it seems that the pre-WWII conventional wisdom was "1 pulley, straight chain, reduce friction" and thousands of derailleur variants were developed to these guidelines. However, once Campagnolo introduced the modern-looking 2-pulley parallelogram Gran Sport in 1950, they all instantly became obsolete. The added friction seemed to outweigh all the other myriad problems of the predecessors.

John Natiw said...

This is why I read you Dave. The very definition of "outside the box".

Anonymous said...

"If the chain is to remain out of alignment, how about a chain with spherical rollers at each joint so it will run out of alignment without the friction of the side plates."

It would run smooth, but it wouldn't shift. A laterally stiff chain is required, front and rear, when shifting from a large cog to a smaller one.

Anonymous said...

Shimano FFS was bound to fail because people subconciously said "For Fuck's Sake!" every time they saw it.

Affect = verb
Effect = noun

Anonymous said...

I know next to nothing about improving shifting technology, so here's my barely useful contribution:

Effect = noun AND verb

As in cause and effect (noun) or effect a resolution to this dispute (verb meaning "cause to come into being")

Affect = noun AND verb

Affect as noun means the "manifestation of an emotion"

Affect as a verb usually means "to put on a show of some emotion" but it can also mean "to produce an effect upon"

Moral? English is a pain in the ass.

lemmiwinks said...

Indeed it is time for a re-think of the dérailleur. Get rid of it and use an internally geared hub ;-)

I'm a happy hub gear rider (ancient Shimano 3S for the curious and yes, I know about their reputation for breaking, in fact I broke it on Monday but fortunately I had a spare to rebuild the internals with :-)).

I care little of friction issues! I ride, in no particular order, for transport, health and pleasure. Sometimes all three at once. My bike has 27" wheels with 1"1/4 tyres on it (block tread on the rear which sports a steel rim no less!)

I actually have both front and rear dérailleurs on my bike as I added a mountain bike triple obtained cheaply to give me a spread of 9 gears. The rear dérailleur serves only as a chain tensioner, and while I probably could have used something like a Surly Singleator I had the dérailleur on hand.

I use the original friction shifters for changing both the hub and the front chainrings. It's a lugged steel mass produced frame and I love it, though with it's mish-mash of MTB cranks, bars, hybrid forks, road frame and wheels, and sturdy rear rack I'm sure Dave would hate it :-)

I lean towards the "retro grouch" end of the cycling spectrum (though I've bought a cartridge bottom bracket) which is probably why I enjoy this blog.

Dave comes across as absolutely certain of his being correct in every post. Whether he is correct or not, whether you agree with him or not, you gotta admire that confidence :-)

Anonymous said...

And then there's this:


mark worden said...

Visualize a gear pack where each cog could telescope inside the next larger cog so that the cog in use is always in the same alignment position. If you were on the highest gear, smallest cog, all 9 cogs would be visible, and in the lowest gear, largest cog only that cog would be visible. All the others would have tucked into the cog next larger in size.

The movement of the chain would not be in and out, but up and down, always in the same alignment. The front cranks could also use the same technology. The alignment would always be perfect. Now figure out a derailler system front and rear that moves the chain up and down smoothly with the movement of the cogs/rings in and out.

I bet someone a lot smarter than I has already thought of something like that and decided that it was not a good idea--but who knows.

Anonymous said...

What level of friction are we really talking about here as compared to all other forms of resistance to forward motion on a bike? I've always heard that the chain drive system of a bike is incredibly efficient.
Bladed spokes make a huge percentage improvement in aerodynamics over a round spoke, but when analyzed as a percentage of the total resistance on a rider the overall effect is pretty much imperceptible (measurable but insignificant).
Are we talking about the same supposed improvement as going to the new fangled ceramic bearings? I somehow doubt that Dave would be in favor of going to ceramic bearings if they would have the same effect in reducing friction as his "improved derailleur" concept.
We can debate and argue all we want here, but will someone with actual knowledge of the topic weigh in and tell us what level of friction we are dealing with in a dual pulley derailleur system?

Anonymous said...

Hey Dave, I like your blog and have it added to my links of bike blogs. I'd love the entire blog submission so I can post it up on the front page. The submission form is on the left on my site...

VintageSpin said...

There are no scientific studies done on drag and friction in the derailleur system (at least not peer reviewed and published, only independent). Apparently there isn’t a good reason to do one, and who’s going to pay for it, the derailleur manufacturers?
Although the chain and sprocket drive is one of the most efficient systems of power transfer yet invented, it is the introduction of the derailleur that causes more drag and friction, and there is no way around it…yet.
It is just that aerodynamics is stalemated; weight is moot; what has been left out? Friction, drag on the mechanical parts of the bike. Some of this has not been addressed by manufacturers, thus the point of this blog.
It isn’t up to the riders to figure this out-we’re too busy riding what we’re sold.

Anonymous said...

the book "bicycling science" does go into detail about chains/friction and it turns out to be minimal in terms of other hinderances to forward movement ie weight/drag.

They say even rusty chains are nearly 90% efficient, and clean/straight ones near 98%.

Once the divici cvt hub looses enough weight to be competitive i see that as the future of geared drive systems. Great blog.

Anonymous said...

Observed trials bikes often use front freewheel mechanisms with fixed-cluster rear ends. The use is to both easily increase engagement points with high-end front freewheels, and to spare the rear wheel from a complex system that endures tremendous impact from the best riders! Though often, gear clusters are simply cosmetic to oblige the rules, and are not shifted.