bt

dude you are spreading bad information.

the Di2 FD works flawlessly and is easy to setup.

And if you know of a specific bug with the first gen di2 shifters find out what is was before you slander the product that others may be interested in buying.

spectastic

oh yea, I'm sure you'll never run into any problems with your electronic groupset, like ever.

grolby

Citation needed. Short-range low-energy wireless is very reliable these days. The biggest consumer problem with SRAM's new system probably isn't going to be wireless reliability, it's going to be keeping on top of charging the separate FD and RD batteries (though they have a charger for both to plug into that should make it easy to remember). I don't think batteries for the shifters will be too much of a problem - I would guess that they'll run a very long time on watch cell batteries or something like that. It's the motors that suck down basically all the power in an electronic shifting system.

The motors are also probably one of the biggest reasons that it's taking so long for the technology to trickle down. Sourcing electric motors that are small, powerful and precise enough, especially for front shifting, has to be expensive.

Anyway, there's a lot of sturm und drang about electronic shifting being unnecessary or somehow destroying the purity of a bicycle. I don't really buy that. It's all unnecessary, so calling Di2 unnecessary while riding around on your 10 or 11-speed groupset is a bit ridiculous. And it's just incorrect to see the shifting system as having anything to do with the human-powered nature of the bicycle. Shift cables functionally don't transmit energy, they transmit information. You can send that information with a mechanical system, or you can send it more reliably with bits. This just isn't a big deal.

There do seem to be some issues with electronic shifting. The big one is that dropping a chain seems to be much more of a showstopper. With mechanicals, you can usually finesse the chain back onto the rings with the shifter. That doesn't seem to work with electronics most of the time.

justinzane

Ok, I can see why this might not be "Electronics". But it is applicable to any multispeed bike -- road, mountian, TT, utility, etc. Part of my point is that while the current products are targeted at high-end racers and wannabes, there is no non-marketing reason for this to be so. In fact, most serious road cyclists can probably use friction bar-ends without losing much time or having any difficulty shifting.

Most casual cyclists, on the other hand, would freak out with friction barcons. (I know, I had my wife try my SunTours in both friction and indexed mode. She hated them.) They, on the other hand, would be very well served by having a simple upshift and downshift button on each side. Even more so a single up/down button with the logic in the microcontroller to do double shifts as needed to move through the entire 28-28 through 48-13 range (or whatever the specifics of their crankset/cassette are).

I'm a total noob; yet I find that even after only maybe 50 miles, my right pinky (finger) instinctively knows just how hard to pull from one gear to the next. But, I put some effort into carefully adjusting the derailleur and learning the feel of the shifting. With many Americans, and probably others, unwilling to put effort into learning much, preferring to buy devices that "just work", this seems ideal more for the casual cyclists rather than the serious ones.

justinzane

I am using "shift-by-wire" in analogy to "fly-by-wire" in aircraft control systems. Even though some planes may use fibre optic data transmission or RF data transmission, the general use is to indicate that the pilots' controls are not directly mechanically connected to surfaces on the plane. Same deal here.

Sorry, nit removal.

justinzane

I use Suntour Barcons now. I am not asking because I need electronic shifting. I am just curious about this area of cycling development. I'm the kind who likes to learn about lots more than he will ever actually buy.

gsa103

Wireless is very reliable, but its a long way from the 99.9% reliability of mechanic systems. I use a Garmin 510 paired to my phone with HRM and cadence. Periodically, something just doesn't synchronize correctly, sometimes its low batteries, other times a software update or something reset etc. My experience is that no wireless system is ever more reliable than its comparable wired system. Consider Wifi vs Ethernet, cell phones vs land lines, wireless never works better but its frequently more convenient. I imagine most people have had similar experiences. If my Garmin doesn't work I can still ride, if my RD doesn't work, well that's a problem.

The other thing is that I don't see the point of wireless. Its not like you're going to be removing the shifters or derailleur constantly (unlike a Garmin). So basically you're adding two batteries and a potential interference link to get rid of the hassle of running a few wires when you build a bike?

justinzane

I'm sorry, but this is an area that I actually have some professional and practical experience in. And relatives who have vastly greater experience, making things like guidance controls for a famous point defense missile system. The technical difficulty in getting a reliably working system are extremely minimal. Consider for a moment, things like DVD players. They move a tiny laser diode, lens and sensor across the surface of a rapidly spinning disk with sub-micrometer precision and the ability to handle odd orientations, systemic motion, etc. This is done using the same microcontroller plus servomotor plus position sensor combination that is useful for shifting derailleurs.

In an age where my two year old phone is a more powerful computer than the ancient (1995!) mail server that I still keep running for sentimental reasons, things like developing lightweight, space efficient, water resistant, shock resistant batteries, integrated circuits, micromotors, etc. are old news. The amount of logic needed to have an electronic shifter adjust itself on installation to just about anything from 5-11 speeds in the back and 2-3 in the front with any reasonable sized cogs is minimal.

I have the equipment to do something like this DIY-electronic-shifter, but not anything better fit to a bike. Anyone with the capital to contract with a manufacturer in China, Taiwan, RoK, etc. would be able to make something much smaller and better fit. I also could write the code for an AVR or ARM microcontroller to handle the initial setup and continuous positional adjustments needed to ensure quality shifting. And, I am not an excellent coder by any stretch of the imagination.

The only area that might require significant R&D is setting up the packaging and adjusting timings for serious competetive cyclists -- which is the market that has already been met. Time trialers and triathlonists may care about grams of drag; the tourer, urban commuter or recreational cruiser likely has already done so much to increase weight and drag with panniers, baskets, flappy clothing, and upright position that any impact from an affordable electronic shifting kit would be truly trivial.

Unless someone has specific, technical knowledge contradicting this position, I'm really pretty sure that engineering is not the issue.

grolby

I'm not an expert on wireless technology, but what a wireless shifting system has to do is much simpler than any of the counterexamples you're giving. There are plenty of things I've done with wireless pairings that work essentially flawlessly - I'm typing this on a Bluetooth keyboard right now, and it never drops unless the batteries run down. My coded HRM works perfectly as long as I'm in range of the receiver. And SRAM builds and controls every piece that needs to pair up and work together. Do you REALLY believe that SRAM would even consider releasing a wireless shifting system that wasn't absolutely rock-solid? Hey, it could go wrong for them... but I don't think they're going to screw up that badly. The battery thing really is the most likely source of problems.

I'm not defending wireless from your other criticisms, requiring multiple batteries isn't really ideal. But there's no reason a wireless system can't be reliable enough. It absolutely can. If that's where SRAM's system falls down, it will be due to incompetence, not because the technology isn't reliable.

grolby

justinzane, you are displaying classic Out-of-Field Engineer Syndrome. "It's not that hard, because X, Y and Z are not that hard in the domains that I am familiar with." Well, maybe it seems easy, but the evidence suggests quite strongly that it is not that simple. Shimano and Campagnolo had electronic systems in testing for the better part of a decade before finally releasing a consumer product. They clearly did spend millions of dollars in R&D. SRAM had a late start, but has certainly been working on their still-unreleased system for at least four years. What are they spending all that money and time on? I don't know. What's challenging about making an electronic shifting system? Again, I don't know. But the facts on the ground are that they do need to spend time and money, and coming out with a good product for the consumer market is difficult. My lack of specific technical knowledge about why it's difficult isn't relevant. The barest examination of the facts should make it abundantly clear that it's hard to make an electronic shifting system.

justinzane

What I'm trying to get at is that this is not, from a purely engineering and manufacturing point of view, a very expensive thing. Sure, I would be terrified tossing a $10,000 bike on the back of my car or riding in a crit. But, I would never spend $10k on a bike in the first place. I'm wondering why, in the unique world of cycling components, this particular piece of equipment has not become more common and more affordable, especially since there are plenty of manufacturers outside the bicycle trade with the ability to produce electronic shifting systems affordably.

I'm sure that I'll be dead before there is a Campagnolo Sconto groupset let alone such a thing with EPS. On the other hand, while one can drop $10,000 on a Felt AR FRD, it is also possible to pick up a Schwinn Prelude for under $300. Same goes for bars, cassettes, bottle cages, saddles, etc. I'm sure that at some point in the next hundred years that electronic shifting will be available in the whole rage from WalMart $19.98 junk through Campy's Super Record gorgeous, perfectly performing, for the real-pros-with-sponsors options.

I've spent many years in the IT field; and I understand the economics and pricing fairly well. Being new to cycling, I'm kinda mystified by some of the things I see. I'm curious to learn from you and everyone else how things work here, not simply wanting to buy something now.

Thanks.

Brian Ratliff

1) Wireless is not necessarily unreliable, but cross-talk has to be an issue when you are in a field of 100+ riders. HRM and other wireless data devices still experience cross-talk on occasion; this can be very bad when talking about shifting. I am not saying the problems are insurmountable; SRAM apparently is trying; but for a problem to be solved incentives have to be going the right direction and both Shimano and Campagnolo have decided to use wiring harnesses for the foreseeable future. Those companies have their share of bright engineers; I imagine they've gone through the options.

2) The chief problem that electronic shifting has always had coming to market is how well the mechanical systems work. Electronic shifting will always have to compete with mechanical and if Shimano or Campy just decides to eliminate mechanical systems from their catalog, the technology is mature enough that other companies will fill that void almost immediately. Mechanical systems are extremely profitable for Shimano and Campagnolo; I can't imagine that a $400 shifter set costs more than $20 to manufacture at their volume. Oddly, the cost of batteries might be the chief problem. Only the manufacturers know exactly how much it costs to built the shifters; you won't see electronic systems taking over the market until their manufacturing costs drop lower than mechanical systems.

3) Bike component companies also have the problem that their "volume" looks a lot more like "one-ish-off" production from the point of view of many industries. You won't see Shimano influencing the price of batteries like you see Apple influencing the price of batteries. Bike components are not commodities; the price they sell at is driven by demand, not manufacturing cost. Also, bike components that we tend to talk about here in the road cycling forum are basically luxury goods from an economics standpoint (the guys making money riding their bikes are typically not on the same price structure for components as the weekend warriors), which tend to have perverse pricing structures (more expensive, to a point, might lead to an increase in demand rather than your typical commodity pricing behavior)

justinzane

Thoroughly agree on that one. I'm not exactly sure why wireless is useful, though I've personally got no problems routing brake and shift cables. Perhaps SRAM is going for the doesn't-own-a-screwdriver crowd. Wireless seems useful if one wants to have realtime datalogging to a phone, though.

My hunch is that it is only expensive for the truly high end where a few milliseconds slower shifting is noticeable and perhaps critical. For the lower ends of the spectrum, gearing should allow much lower cost motors and much more accuracy in positioning. Using a servomotor with say a 1:1 drive ratio will allow extremely quick shifts with a strong enough motor and battery; using one with perhaps a 10:1 ratio will be slower but able to position more accurately and with lower power.

Basic ~$50 retail hobby motors can do a revolution in 80ms with 80kg*cm torque. IIRC, deraileur return spring tension is somewhere within the order of 10-20 kg*cm. That should mean that, if ones buys motors in quantity, a motor should cost less than $20.00 to be sufficient. Please correct me if I've got the numbers really wrong. I'd love to hear from someone who has specific knowledge on this point!

Agree again. If one wants purity, I'm quite sure that steel frames with fixed drivetrains will be available for far longer than I am around.

That I'd not thought of. Though a decently designed and setup system should not over/undershift at all; I can see how that could be more of an issue for those who ride off-road. I've only ever dropped a chain because of misadjusted limit screws; but my experience could fit on the head of a pin.

justinzane

Apparently I am as unclear in text as in voice... I was hoping to make the point that reliability is not a problem, quite agreeing with you.

grolby

No worries, I wasn't referring to you, I was referring to gsa103.

As for everything else: you could email someone at Shimano and ask them why Di2 is so expensive and why it is taking so long to trickle down through their product lines. I doubt they would explain it to you. And no one here has any specific knowledge of the process Shimano took in designing Di2, though if they did, they certainly couldn't share it with you. All you're going to get is speculation.

Lots of people come into bikes from their own domains with the assumption that bicycles are easy. They always end up surprised to learn that bikes are not easy at all. Some of the "not easy" is on the engineering side. Some of it is because people often fail to understand that bikes are products, and taking something from an idea to a finished product is really challenging and requires lots more money and expertise than most people think. It's that much worse when you consider how small the bicycle components aftermarket really is. PowerGrips boxes currently boast "50,000 sold!" Let that sink in: 50,000. PowerGrips is a niche product in a niche industry, but that's just a tiny, tiny number. But it's a success story in this business. And then there's the fact that the bicycle industry is incredibly competitive, particularly when you consider that the absolute financial stakes in the industry are so small. There aren't a lot of millionaire bike company founders out there.

Now, if you want speculation as to why Di2 is still an exclusively high-end, niche product, I can give you some but it won't be that exciting. Basically, it's probably not feasible for Shimano to produce a Di2 product line below Ultegra level, for a combination of techical and marketing reasons. Are they capable of producing a 105-level electronic group? Almost certainly, but doesn't mean it's a feasible product.

For one thing, as you go downmarket, the products have to be less nice than the high-end equivalent. That's partly to reduce costs, but also for straightforward marketing reasons. The more expensive product has to seem worth the price premium. But, on the other hand, the product still needs to perform well, and Shimano is very sensitive to that. Their mid-range and low-end products work great, and have for decades. There may not be much room to bring the Di2 gear downmarket without hurting its function. That might not be true, but it's possible.

Another issue is that Di2 is already sort of a premium upgrade to the high-end Dura-Ace and Ultegra lines, probably because it pretty much has to be. That's fine at the high-end, but a premium version of 105 or Tiagra would probably be difficult to market even if it is feasible to produce. That could change, but it will take a long time. I would guess that, at some point, Dura-Ace will be available only in an electronic version. Maybe then Di2 will move downmarket. But I wouldn't be willing to put a lot of money on that bet.

Not to mention that even a lower-end group with trickle-down technology requires a design process. So there's that additional cost to consider as well. Is it worthwhile to invest that much more time in 105? Maybe not.

As for pure technical issues, it's true that electronics aren't the issue. Those parts improve so quickly that it probably isn't the issue. But batteries, electric motors and manufacturing technology are on a much slower roadmap, and the bike industry is far too small to be able to exert much influence over how that technology improves or where it goes. So it'll take a while.

Personally, I think it's probably in Shimano's interest to get electronic shifting almost everywhere. And I bet they think so, too - Alfine Di2 was a pretty clear signal that they don't think electronic shifting is just a racing technology. But realistically, it will take a very long time for it to be feasible for them to do it.

RollCNY

It may just be me, but it seems like you just insulted a fairly large group of people in one fell swoop. By this argument, US residents should simply give up on cycling all together, because if they haven't learned to shift yet, how are electronics going to help?

hamster

Before you can produce electronic shifting systems, you need to know how to produce mechanical shifting systems. There are probably fewer than ten companies in the entire world with existing designs and product lines for the production of rear derailleurs. Entry of the new players is blocked because low-level mechanical components are cheap and it's hard to beat Shimano and SRAM in weight game.

Weight is a major consideration. The derailleur needs to apply a lot of force to the chain in order to shift quickly and reliably under load. You need to figure out how to do this without a proportionally large and heavy servo. Shimano gets the job done in the 6870 group by adding 75 g in front and 75 g in the rear. It's probably harder than you think.

There were several attempts to do electronic shifting before Di2. They all failed because they were too finicky or too expensive. For example, one, Mavic Mektronic, was wireless and had a tendency to stop working near police speed traps.

justinzane

I'm good at insulting people; I can do it without even trying!

The observation was that many casual cyclists who started with indexed grip shifts on their department store bike or their entry-level LBS bike are not comfortable with friction shifting and are not interested in spending the few hours/miles necessary to learn that skill. Just like most cyclists have never swapped their cassette from whatever came with their bike when the ride in different terrain, or have never bothered to figure out what pressure is appropriate for their weight and tires and simply pump up the the labelled max.

How many times have you seen casual cyclists riding on small-small or big-big, rubbing their crossed chain against their untrimmed deraileur? The may well have little interest in how to properly shift; but if their bike presented them with simple up-down buttons and handled the logic of moving from granny-granny up optimally, they would be happier riding.

I'm talking about the people to whom visiting bikeforums.net is never a thought. They buy a bike and jump on and expect it to "just work". Everyone has different interests and priorities. For those who do not really care to learn how to most efficiently use derailleur gears, programmable logic may be of noticeable benefit.

justinzane

Thank you! Very good point about the small size of the whole market limiting the amount of commoditization that is possible. Though there are probably over a billion cyclists in India and China, they are unfortunately not in a position to afford even $50.00 electronic shifters and what we might consider a uselessly cheap groupset. So, in a world where 50k is a huge success, the ability to move manufacturing costs through volume just is not anywhere near what I expect. Sensible.

And, with the mention of the Alfine Di2, you remind me that one of the more utilitarian things might be simply to avoid derailleurs for casual use. The internal gear market seems a bit overpriced, but as quality automated machining equipment drops in price, it seems that those will slide down into the entry level range. There, there is just no getting around the fact that very tight tolerance with high-spec alloys are critical. Have to think that through some more.

justinzane

I'm just not sure about that. While full integration might be nice; having a servomotor and spool pull a short piece of shift cable instead of my fingers and a spool seems to have no requirements on knowledge of deraileur manufacturing. Additionally, not making a fully integrated electronic groupset gives the huge advantage of allowing one to use an electronic shifter with absolutely any derailleur, cassette, crankset, etc. Basically, think of friction shifters driven by a motor not fingers.

justinzane

You are certainly right. Unfortunately that conspicuous consumption type logic always makes me absolutely crazy!

Brian Ratliff

Why? This isn't some conscious behavior on the part of a massive group of people. This is the sum total of all our behaviors; we all buy stuff and this is how the pricing works out. It's not a logic, it's the natural behavior of the system. You even see it in nature with females and males looking for mates. Peacocks have their shiny feathers. Elk have their oversized racks. This is the nature of desirability put into the context of product markets.

Brian Ratliff

It requires concepts of indexing to align the chain and cog, and it requires a locking mechanism (unless you like running your servo in a static position, sucking up battery life), either explicitly with a brake or clutch, or implicitly with a non-backdrivable gearset between the servomotor and the spool. This indexing requirement is rather difficult; I believe the solution settled on is a very basic machine vision system with sensors which tell the microcontroller when the chain is aligned. You can't just count revolutions on the servomotor either, unless you enjoy adjusting your system by hand nearly constantly.

hamster

This is an interesting approach, but it's guaranteed to be heavier than a fully integrated group, and, unless you do it right, you may lose the benefits of having electronic shifting in the first place.

RollCNY

I get your point, but I don't agree that electronic shifting makes that any easier. I have found that many riders don't understand why you would want to shift, so electronic or mechanical is meaningless, and many of them shift fine once they talk to someone about it, and have the rationale explained. Look how many "I keep dropping my chain" threads there are on BF, that get resolved once they talk it through.

Electronic shifting is not simpler than mechanical. I ride with several guys with Di2 who are fairly savy, yet they still make frequent LBS trips to keep the system functioning quietly. My Sora group set is quieter.

If you really want to aid new cyclists, come up with a self-adjusting seat post. Most riders I see have their saddles way too low.