One of the virtues of electric vehicles is their inherent flexibility. Electricity is a very nice form to move energy around, and simple electric machines can be configured act as both motors and generators, with reasonable efficiency, size, cost, and reliability. As electricity becomes more integrated into vehicles, it will allow different design choices to be made. Here's one:

Imagine an electric transmission bicycle. The foot pedals power a generator, and the wheel is driven by an electric motor. This is what a car guy would call a series hybrid, which in a car would mean an internal combustion engine driven generator, and electric motors for the wheels. The advantage in a car is that the ICE can be tuned for maximum efficiency, and electric motors are very efficient compared to ICEs which need to function under varying loads. The series hybrid dominates in rail transportation, where it's known as the diesel-electric locomotive. If you consider the physical layout of a recumbent bicycle, one advantage is obvious: most recumbents need a long, ungainly chain drive, which an electric transmission could replace. There are more advantages: no fussing with gear changes, a small battery would allow load averaging, so you could keep pedalling at a stop light then launch using stored power, regen braking, use as a stationary exercycle, etc. These have actually been built, by Andreas Fuchs, who later started Autork to commercialize similar concepts. The disadvantages are pretty obvious too: 70-80% is fairly good efficiency for a motor or generator, 90% is excellent, and the electric transmission pays twice, once in converting mechanical power to electric, and again to convert it back. A well-adjusted chain drive, on the other hand, is around 98% efficient. Since bicycles are low power, and the only way to get more power is by changing the operator, efficiency is very important.

Okay, fine, so maybe that's not such a good idea. What else can you do with electricity? Two wheel drive is another application of the design flexibility. All-wheel drive is becoming increasingly demanded on cars and trucks, for reasons of handling and traction, but is even more important on a two wheeled vehicle, which starts out less stable. Unfortunately, it's also considerably more difficult to implement on a two wheel vehicle, because there's no easy way to arrange a transmission. It has been done however, both with a mechanical transmission in the Rokon motorcycle and hydraulic systems commercialized by Yamaha (the 2-Trac) and KTM. The disadvantages of these choices are fairly clear, though: the Rokon has no suspension (a serious failing on an off-road machine!) and turns the power off when you turn the wheel, and if you thought electric drive had efficiency troubles, hydraulic is far worse. That's acceptable for an overpowered motorcycle, like the high end dirt bikes KTM and Yamaha sell to professional racers, but not for a bicycle.

So, how do you get both efficiency, full suspension and AWD in a bicycle? Simple enough. It looks like a regular bicycle, with a chain driving the rear wheel. The rear wheel, however, has a hub generator, similar to the hub motors put on power-assist bikes and the hub generators used in high end generator driven bike lights such as those sold by Peter White. The front wheel has a hub motor, too, and in between them is a controller which monitors the relative slip. When the rear wheel starts slipping, the controller takes power from the rear and drives the front until they are the same (or less than 5% rear relative slip, eg). The back emf on the rear acts as a kind of ABS, braking the rear when it starts to slip. You can't slide the tires unless you're sliding them both at the same time, and there really is no traction to be had anywhere. When you have adequate rear traction, though, which will be most of the time, power gets from the pedals to the road through a highly efficient chain, there will be no load on the generator, and careful design should be able ensure low residual drag. If you want front wheel drive only, just transfer all your weight to the front and pedal hard. As compared to a power-assist bike, there is no battery, which encompasses most of the weight and trouble of a power-assist system, as well as a good deal of the cost, there is no limit on range, and the two hub electrical machines need only be fairly low power, 100-200 W perhaps, because each only has to handle half of what the rider can put out. As a bonus, even high-power lights are a cheap and easy add-on. While not free, the added wholesale cost should be less than $500, perhaps much less, which would be entirely acceptable for a high-end mountain bike.

I'd thought I was fairly well prepared. I'd had a month or so after ordering before the parts arrived, so I'd gotten all the greases and lubes I thought I might want, and had taken everything apart, ready to put back together with the new bits. The parts still hadn't arrived, so I built a lighting system for it out of Chinese driving lights from Wal-Mart ($19.93) mounted with grounding clamps ($2.49 each). Dual MR11 bulbs, 12 V 20 W, wired in series for the 24V battery. It's a lot of light for a bicycle, but it's small compared with the motor, and I want oncoming traffic to not only notice I'm there, but also wonder what's coming at them. By the time I'd taken the old motor completely apart, just to contemplate the interior and wonder what might be wrong with it, the new stuff arrived.

I had a choice to make first off. I'd ordered two things: new motor, and new gearing kit. The idea of the change in gearing is to trade off low end torque for higher speed, which I thought would be a good idea for commuting in traffic. So did I want to try just bolting the new motor on first, an easy job, and see if everything worked and go from there, or did I want to do everything at once? I'm an optimist, I chose the latter. The new pinion fit on the old shaft, I got the order of washers, ring clips, and spacers correct after only five or six tries and repeated consultation of my notes and the kit instructions, everything fit. I worked slowly and carefully, greasing things nicely, torquing the bolts firmly in sequence, sure that I wasn't going to be inside again for a long time. One last piece: the chain.

The new gear is bigger than the old gear, so my old chain is too short. No problem, the kit came with a tiny piece of chain (one and a half links) to extend it. Put them on, put on the master links, and we're done! Well, it's kind of slack, actually. Really slack, it sags more than an inch in the middle. And yes, a slack chain can carry power, I turned it on just to see if the motor worked and could spin the rear wheel (it did), but it flapped around everywhere. Not really acceptable. A little thought and consulting the instructions showed that the extension was not supposed to be the right length, rather it was just a converter, so that you could add an odd number of holes to the chain. You're supposed to remove links from the old chain, add this piece plus a second master link to make it work. Okay, so how do I do that? And without damaging the rest of the chain? The instructions were mum, just referenced more instructions on the seller's web site. I tried prying a link off with a screwdriver blade, but I couldn't even get it in. I came in, put the kids to bed, and looked to the web instructions for enlightenment.

After treatises on the mathematics of chain length adjustment, I found the part I was looking for, how to remove links. It said, more or less, "There are a variety of ways to do this, and the method selected should depend on the skills of the mechanic and the capabilities of the shop." Oh, c'mon! I have no skills, and my shop has no capabilities. Couldn't you at least mention one of them, so I could go try and aquire some skills and capabilities?

At this point, it's getting late, and I'd like a functional bike. I decide to punt on the new gear, and go with my old chain (hope I didn't damage it too much trying to rip that link off!). I strip everything apart, put the old pinion on, put it all back together, feed the chain through. Now just put on the master link, and I'm done.

At this point, I realize that while I've taken it apart a number of times, I've never put the chain back on. The chain is tight, and one side of the chain has about 2mm of clearance away from a plate. To avoid trying to assemble the clip etc there, I'm certain that I need to get the main part of the link inserted from that side. Nowhere works. I succeeded before because the chain was loose. No way of holding the link and chain works. No way to loosen anything. I'm starting to wonder if it was ever together.

Now the inevitable happens. I drop the link. Ting! Instant fear. I'm sure another one could be aquired, somehow, but I don't know how, and if I pick it up immediately it'll be right here... or here... Jeez, wasn't I going to clean this place up before I started? After a long search, reality begins to dawn. I didn't hear it hit the ground because it lodged itself inside the housing of the pinion gear I'd just assembled. With squinting and a flashlight, I even convince myself I see it, but poking, shaking, and spinning the gear don't dislodge it. There's no hope. I have to take it all apart again to get the master out.

At this point, I'm no longer working slowly and carefully. But at least I'm getting lots of practice.

Unbolt everything, unscrew everything, pull it out and... not there. Nope, not there. Hm. Good to know, I guess, at least I won't be wondering whether it's in there. And back together it goes, after having eaten yet more of my life.

So what now? Well, there is one more master link, the one I need for the kit. The kit's useless at the moment, so let's use that one. And forget about putting the link in from the low clearance side, put it in from the easy side, and getting the plate and clip on from the hard side is, while not easy, at least possible. And it works! Amazing.