kyselad

I actually agree and still run triples. What you describe is what inspired my post: it’s easier to downshift to a smaller chainring under load vs shifting to a larger cog in the rear. But a triple is a lot different from 9+ gears over a wide range, and I’m realizing there are real practical limitations to achieving up front the same gearing options possible on a rear cluster.

PeteHski

The basic physics of moving the chain under tension. It’s just inherently a lot smoother shifting at the rear, where everything is spinning faster with lower torque. Also no chance of dropping the chain on a rear shift.

I think the front derailleur is fundamentally inferior to the rear derailleur and hence why it is slowly being phased out in favour of wider range cassettes. The front mech is already history for mountain bikes, where it’s fundamental flaws were most obvious. On a road bike front shifting is much less demanding and so it’s not really a problem. But I don’t want to swap my 2x12 for a 12x2 thank you!

big john

While it may be difficult to shift to larger cogs under load, it really shouldn't come up very often. If you're paying attention you get in a lower gear before the steep section, especially if it's so steep that you need a smaller ring.

I'm not trying to sound "elitest" but gear selection is part of learning to ride hills. Sure, I've found myself in a too high gear on climbs but it certainly doesn't rise to the level of needing a different type of drivetrain to address it.

PeteHski

I totally agree and yet on group road rides I often see experienced riders get caught out in a high gear and then try to crunch it up into a lower gear under heavy load. Personally, I never get into that situation, but maybe because I ride mountain bikes a lot on steep, hilly technical trails (1x12 drivetrain) and even our local roads are very lumpy. I just shift down before it's too late!

njkayaker

Rear shifting is faster and more reliable because it isn't under tension and doesn't care too much about load. The basic idea is that one should be doing rear shifts much more frequently than front shifts.

njkayaker

The OP isn't talking about triples.

The OP is talking about "many" gears in the front. Given that the current norm for the rear is 10 to 12, you really have to address why there aren't more than 3 used in the front.

Except for really odd examples, 3 appears to be the practical limit.

Note that many gears in the front might require a bigger derailleur, which there really isn't that much room for.

It's not fashion.

Triple are harder (more complicated) to use and there's the issue of trimming. That you and some others (like me) have no real problems with them, doesn't mean that other people don't.

With fewer gears in the back that had a narrow range, triples were much more necessary/useful than they are now. Also, with standard double cranks, bicycles were routinely geared too high. Triples are a way of "fixing" standard double cranks.

njkayaker

There are too many cyclists that don't really know about what you are talking about here.

njkayaker

There's a technique for this too. If one is clear on how things work, it's even possible to do clean front shifts while standing.

njkayaker

The front isn't easier than the rear. Keep in mind that the rear/front derailleur stuff is really old. If it wasn't the better way of doing things, it wouldn't be done that way.

The rear derailleur moves the chain where it isn't being pulled. This makes the shifting faster. And you can shift the rear under loads that do not work well in the front.

The Pinion internal front shifting shows there are alternatives to derailleurs but it's also (likely) much more expensive.

tyrion

An RD uses 2 jockey wheels to guide the chain into precisely the correct plane. An FD uses a cage (which is less precise) to move the chain to a plane that is close enough to change chainrings. To make an FD that uses 2 jockey wheels to guide the chain would be difficult because it would be guiding the part of the chain that's under driving tension - makes no sense when it's so much easier to do that at the RD on the part of the chain that's under much less tension.

Darth Lefty 

Front shifting stinks because it's done on the tension side.

1xN is still not 100% compelling for roadies.

Mountain bike suspension is aided greatly by 1xn in a couple of ways. That's why MTB bought into it so fast and hard. The suspension main pivot can have its bearings much further apart (Boost and 1x add up to like an inch wider for the shaft), and the chain line by the main pivot is consistent to dial out the pedal bob. It happened at the same time as the plus-tire fad, and that helped to keep the chain from rubbing on the tire, and on hardtails keeping the chain ring small for a chainstay yoke and a short rear center

njkayaker

The rear derailleur is fairly gentle. The front derailleur is much more brutish. The difference is not that obscure.

HTupolev

The main reason things trended that way historically is that there's more happening in the front competing for lateral space: the crank arms and riders' legs need to be there, and forcing them far outboard isn't ergonomically optimal.
Furthermore, the combination of geometry (influencing practical wheel sizes), the low-rpm high-torquer characteristics of the human body, and the mechanical convenience of half-inch chain pitch, all come together to require that the front sprocket wheels be pretty large. So having a lot of them up front results in an alarming amount of sprocket happening.

The lack of tension on the lower run of the chain also helps make rear shifting lighter and more consistent, so it makes sense to set up a drivetrain so that most shifts occur there. However:
1-I don't think this point is very helpful in explaining why things were historically set up the way they were, because rear shifting used to be more sensitive to upper chain tension than it is today. Prior to shift gates, chains basically had to be ripped up and off a cog before sliding over to the next, and this could get baulky under pedaling force. (And even today, rear shifting works better if you ease up while the chain moves.)
2-I also think this point gets overstated. In my experience, most of why front derailleurs shift "worse" than rear comes down to what they're being expected to do: huge jumps and huge changes in ratio. On my setups where they're handling 10-tooth jumps and below - even without shift gates - they can be made quite snappy and buttery smooth. This is actually part of why I tend to favor triples over doubles: it allows for smaller front jumps. When my triple setups have shifting issues, it's usually because I'm trying to coax them into doing things that couldn't really be achieved with a double.

mschwett 

but (and this is implicit in your description) the movement of the cage up front is basically shoving the chain over, making contact between the interior of the cage and the moving plates of the chain. it's not a very elegant system. at the rear, the chain is running through the jockey wheels with the contact between lubricated rollers and teeth, and the whole assembly slides without anything contacting the parts of the chain that aren't meant to be contacted. i can imagine that if you really wanted to you could introduce something similar up front, but the tension would make it much, much more difficult to design, right?

mschwett 

on the subject of the current state of the art of the RD and shifting under load, a friend and i were discussing the benefits of SRAM's new t-type "transmission" system which is apparently the best yet at shifting under full load due to a number of related optimizations (chain, hanger, narrow-wide, tooling/ramps). i got to thinking that on my di2 bike, i shift under load a lot more than i thought. although i avoid shifting into a bigger cog while absolutely desperately standing on the thing up a steep grade, in basically any other situation it just shifts nearly instantly without fuss. i cannot say the same for the FD on the same bike, or any bike i've ever ridden!

video showing shifting under fairly constant load, slightly uphill, 250w steady output or so:

https://vimeo.com/manage/videos/844776697/772ba7e8a8

njkayaker

I suppose the chain could be derailed in the front from the bottom (like it is in the rear) but that would be a poor place for a delicate mechanism. (The rear derailleur is exposed too.)

It is what it is (for unavoidable reasons). It might not be fair to say it's "not elegant". They've had many. many years to get this to the best it can be!

mschwett 

would it work from the bottom? seems like you need the chain in the correct position on the leading edge of the chainring, not after the trailing edge?

i agree that overall the system is elegant, combining the driving chain with the gear reduction, essentially getting it for free in terms of efficiency, which is borne out in the 3-5% lower efficiencies of a belt or chain plus a gearbox. the metal to moving metal contact of the FD seems suboptimal.

kyselad

I'm not sure I understand what you mean that the front isn't easier than the rear. Empirically for me, it's easier to downshift to a smaller ring in the front than to downshift to a larger cog in the rear. I do see a number of good reasons why things turned out as they did that have been discussed in the thread. I don't think that seeing the way things developed historically is necessarily solid evidence though -- I know plenty of cases in my field of work where technology is still largely on the wrong track because it's just "how it's always been done."

HTupolev

Besides exposure, the other problem with having the front derailleur below the rings is that you'd need to backpedal to shift. Which people generally don't like.

njkayaker

"Easier" is kind of vague. It is fairly "easy" to drop it into the smaller ring. It's really no less easier to drop the rear onto a smaller cog.

One big problem with front shifting is that it doesn't tolerate shifting under load (with heavy pedal) pressure. The rear is much more tolerant of that.

The reason for that is that the rear messes about with the chain were the chain is not being pulled (there's no force on the chain at the bottom). The front is messing about with the chain when it's normally being pulled. This generally means you have to let up on your pedal pressure (especially, when shifting to the larger ring). Even shifting to the smaller ring work better when you let up on your pedal pressure. It's much easier to drop a chain with shifting with in the front. This is why it's "harder" (overall).

Generally, shifting to the larger thing (front or rear) is working against a spring and shifting to a smaller thing is letting a spring do the work. You are looking at one thing (not overall).

Well, it's necessarily not evidence either. (I didn't use the word "strong".)

As it is, it's cheap, light, flexible, easy to fix, fits in the space it needs to, efficient, and works well (nothing is perfect).

We'd have to see those other cases to judge. There's no indication that the derailleur is the way it is "just how it's always been done".

There are also numerous alternatives to derailleur shifting that are obviously not "how it's always been done". If they were so much better, then they'd be much more popular (they tend to be much heavier and much more expensive).

spelger

i think the chain being fed from the top would not permit that to work.

i wonder if the FD had more rings, say as many as an RD and if the cage were just a bit wider than the chain then shifting from one ring to the other might be as smooth as the RD. im guessing that part of the reason why the RD shifts so much better and faster is that the number of teeth in adjacent cogs is only different by one or two, up front there is generally more than a 10 tooth difference between rings.

this is why i ask about crude, they both push the chain, the FD does it from the outside while the RD does it from the inside.

someone used the word guide but it is still being pushed.

njkayaker

It might not! There might be some way of pushing things up to get it to work.

In any case, the bottom would be a great place to put a derailleur you want to lose!

It's a compromise. (Nothing is "optimal".) As a compromise (being cheap, flexible, and reliable), it seems to be a reasonable one.

In practice, the metal-to-metal contact doesn't seem like a big deal.

njkayaker

No, it wouldn't work.

Yes, this is one of the reasons the rear shifts better. Smaller cog differences shift more cleanly (to larger ones, especially) than larger cog differences.

I don't see being wider helping. Shifting requires only one side or another. The reason it's not wider is to keep the chain in place. Wider would also mean more movement for the smallest and largest sprocket (making these shifts slower).

The idea of the front and rear is that one is for larger changes and the other is for smaller changes. If you make the differences smaller in the front, you'd have to make them bigger in the rear. You are talking about moving an issue (not eliminating it).

"Guide" is a gentle push. The force of the pushing is very different. It's much higher for the front than the rear.

spelger

no, i did mean wider, but just a bit wider than the chain. similar to how the RD's pulley is just a bit narrower than the opening of the chain. right now the cage is much wider than the chain for good reason.

i was just thinking if the drive chain were turned into a 11x2 (opposite of what i have, and the OP's original question) would shifting in the front be better than what it is today and would shifting in the read be worse that what it is. the crudeness flips to the RD but not eliminated from the FD due to...

i am not suggesting this is a good idea though. as all others have said already front shifting under load is not pleasant.

njkayaker

??? It's already wider than "just a bit wider than the chain". What else would you mean other than wider than it already is?

Now, you have a Q-factor problem.

You've provided nothing that would indicate having all those gears in front would "be better". It's more reasonable to expect it to be no different.

Part of the "crudeness" is because the front derailleur is doing work where the chain has tension and the rear is doing work where there isn't tension. Increasing the gears might reduce that (not sure how) but it won't eliminate it.

There are real engineers working on these products. Derailleurs have, quite clearly, stood the test of time. You (probably not an engineer) would have to make a much stronger case that drastic changes to them is a good idea. You are basically assuming that all those people haven't thought about this stuff.

If you don't think it's a good idea, why even suggest it? You haven't addressed the "shifting under load" problem (in any clear/convincing way).