Harold74

The sketch below illustrates my questions. What do we know about derailleur travel paths?

As a follow up question, what do we think the odds are that electronic shifting technology will eventually lead someone to make a universal, "super derailleur" for which the user could programmatically adjust the pull ratio and travel path to suit any setup? I could see that having three useful applications:

1) Vintage bikes.

2) All bikes in a stable when one might want to spend big $$$ on an awesome RD and be able to move it from bike to bike.

3) A custom travel path might improve shifting if it allowed the derailleur to follow the curve of the cogs with greater precision.

The drawback would probably be that such a thing would likely have to be ugly as sin, particularly in the context of vintage bikes.

Litespud

I imagine the RD cage itself describes an approximately horizontal path, more properly a shallow arc. However, the distance between the upper jockey and the sprocket immediately above it is governed by chain length - as the RD moves inward and the chain wraps around successively larger sprockets, the resulting rotation of the cage pulls the lower wheel forward and the upper wheel downward. I think the geometry of the cage and the relative positions of the jockey wheels and the pivot will determine the path described by the upper wheel and the cassette.
iirc Campag claim that their 12sp RDs are designed so that the movement arc closely mirrors the shape of their cassettes

hokiefyd 

The one primary exception to this rule are derailleurs with upper jockey wheels, or guide pulleys, that have an axis coincident with the axis about which the cage rotates. In other words, as the cage rotates downward or forward, the vertical position of the guide pulley remains exactly the same. This is the case with many older SRAM deraillers and older Shimano Shadow style derailleurs designed for 2x/3x drivetrains (for example, the Shimano RD-M3000 and the SRAM X3). Because the ideal path is not a straight line, but rather an arc, I've found these derailleurs hard to tune well, especially with wide range cassettes. If they're adjusted to clear the largest sprocket, the guide pulley will remain pretty far below many of the smaller sprockets (making for slow or difficult indexed shifting). If they're adjusted so the guide pulley tracks closely to the smaller sprockets, the pulley won't be low enough to clear the larger sprockets.

Shimano used to call this a Double Servo panta movement. Ironically, you'll notice they illustrate it with a straight line, but the optimum path is indeed curved.

Harold74

Thanks, I had not considered that. I currently have an old Suntour DR and a new Claris DR on my desk. My impressions of those:

1) The motion of "body" of the derailleur appears to be linear, and would almost have to be based on its construction.

2) The Claris upper jockey does drop down as you suggested.

3) The Suntour upper jockey actually seems to pop UP.

I guess that all is well so long as the various components of motion add up to something approriate.

hokiefyd 

To complicate things, the size and configuration of the cassette really drives the demands of the derailleur. These super wide range cassettes (like 10-51) require a ton of vertical displacement of that guide pulley. And derailleurs designed to work well with these types of cassettes also have larger distances between the guide pulley and the cage axis. The more distance there, the more vertical displacement you can get with cage movement. I think this is a pretty interesting aspect of chain-driven drivetrain design, personally.

Harold74

So it seems to me that, although the effect may be negligible, all of the following would impact the derailleur guide pulley travel path:

1) The B-screw adjustment would impact the linear component of the travel path.

2) For the same chain length, the B-screw adjustment would impact the non-linear component of the travel path.

3) Chain stretch would impact the non-linear component of the travel path.

Harold74

I could have sworn that somebody mentioned that shift "crispness" differences would be negligible on a 7-speed. I can't find that comment now though.

The reason for that is that 7-speed setups have wide, durable chains and larger gaps between cogs, right? Those things lead to more forgiving shifting? Once I run out of 7-speed cassettes, I may attempt to switch the bike to a 9-speed throughout the entire drivetrain, complete with new wheels and all. Hopefully that's a 2030 project.

JohnDThompson 

We already have that technology. It's called "friction shifting."

shelbyfv

That may have been a post I deleted in the thread where you were going on about long cage vs short cage derailleurs. You are correct as to the gist. You will not be able to tell one from another if the cog range is appropriate. If you are determined to go out of spec your experience may vary. Obviously faulty set up or worn out parts won't help.

HTupolev

Pull ratio is fairly easy and probably already doable with the mechanical designs of most, if not all, existing electronic derailleurs. The main hindrance to it is whether manufacturers actually want users to do it.

Programmable travel path is doable, but it would be difficult to justify the considerable increase in mechanical complexity. During Campagnolo's post-Dura-Ace-7400 scramble to stay relevant, one of the weird designs they pushed out allowed the parallelogram angle to be selected between two presets. It made the derailleur heavier and a bit funny-looking, and was generally a failure, solving a problem that not very many people had.
Modern 1x derailleurs do a pretty good job of adapting to different cassettes through the offset between their jockey wheel and the a-pivot. Modern multi-ring derailleurs less so, but they work well enough anyway, and it's not like very many cyclists want to routinely make large drivetrain changes on a multi-chainring bike (having less need to do so theoretically being one of the perks of multi-ring drivetrains).

squirtdad

I think your are really over complicating the derialler thing, but with some interesting thinking

there is clearly no demand for super derailleur or there would be one. Bike companies market what sells...

I can't imagine any one wanting to go to the effort to switch derailleurs between bikes, too much work and hassle. Many people with multiple bikes keep each bike ready to go. I certainly do, meaning each bike has a pump, seat bag with tubes, patch kit, tools, their own lights.

as with all things mechanical one super derailler conflicts with keys to easy success

1. Matching components will work better than mix and match, and certainly will be easier to set up
2. Keeping things in spec or very close minimizes complications

in the past, before the suntour slant parallelogram pantent expired i think there would have been more different paths, but pretty much all current rear deraillers use the slant parallelogram approach

HTupolev

In what way? "Slant parallelogram" simply means that you're not constraining the parallelogram pivots to the fore-aft/up-down plane of the bicycle. It doesn't refer to a single specific geometry. Nearly all derailleurs do use an upper knuckle below and aft of the wheel axle with the parallelogram extending forward, but there's a lot of variation within this theme, and it's simply because it's a practical choice given the derailleur mounting location that we use. For example, the chainstay-mounted Nivex derailleur from the late 1930s had its parallelogram pointing "backwards", because that's what makes sense if you're trying to get the jockey wheel to stay adequately-close to the cogs when your derailleur's mounting point and upper knuckle is in front of the wheel axle. (This geometry also gave generally good jockey wheel positioning throughout the gearing range without a slant, which is why Huret designed the Allvit with a "backwards" parallelogram even though it was designed to mount to a hanger below the wheel axle.)

If you're suggesting something other than a parallelogram, I guess that's an option, but parallelograms have a combination of elegance, durability, and lack of stiction that makes them hard to beat for the purpose.


No, many 1x derailleurs - including pretty much all of SRAM's models and the Campagnolo Ekar - use an unslanted dropped parallelogram. This is usually fine for 1x derailleurs because they can use a high jockey wheel offset to manage chain gap, without fear of front shifting interfering with the jockey wheel position. The benefit is that the parallelogram doesn't get actuated when the chain is bounced by the bicycle going over bumps: the most obvious benefit is that bumps are less likely to cause ghost-shifting, but it also (and arguably more importantly) protects stuff like the shift cable and limit screws from sharp jolts.

Harold74

Was I supposed to do something other than "go on" about the subject of my own thread?

Harold74

To clarify, this thread really has nothing at all to do with my other thread about short cage derailleur compatibility. This one is just about me understanding how things work and what technologies might be on the horizon.

You're probably right. That said, sometimes in business a product has to be brought to market ahead of apparent demand in order for the demand to make itself apparent. In fact, most cycling tech strikes me as stuff that has been brought to market via the racing community before it sparks "regular folk" demand. Nobody really "needs" most of this stuff it seems.

I disagree. The very point of a programmable derailleur would be to produce behavior that perfectly matches specifications. One should be able to match a broad range of components, possibly better than the originally spec'd items would have.

I'm far less of a C&V purist than most but, if someone could make a sexy, shiny, programmable derailleur that would allow indexed shifting on my wife's old 15 speed, I'd be all over that. Plus, I enjoy programming things.