A Taillight for Alan
This has been an ongoing project for a several months. The idea was actually broached to me back in June or July of 09 before I left on my Great Divide trip. My co-worker and friend, Corey Thompson (also happens to be a talented bicycle frame builder), built a frame for a fellow named Alan down in Portland. The bike itself is a Randonneuring bike with all the bells and whistles that you get with a custom frame. Brazed mountings for things like cyclometers, lights and racks; internal wire routing for headlights and taillights. Alan chose the newest Schmidt headlight, the Edelux, for his forward illumination but was at somewhat of a loss for his backward illumination because most commercially made taillights are sinfully ugly, and not fender mountable. After a time he ended up finding an old Luxor taillight on Ebay.
(begin long rant about bicycle lighting)
This Luxor light was made in the dark ages before the invention of semiconductors, so the illumination it provided was sourced from one of those antiquated incandescent light bulbs. My task was to simply take out the old bulb and slap an LED in there to make it bright and wonderful while keeping the stylishly retro flair intact. Simpler said than done. There were several hangups on this project. Perhaps the largest was that the bike was in Portland, and I, being in Olympia, was not in Portland. Therefor, I didn’t have access to the Edelux headlight, and couldn’t do any testing of how exactly this headlight drove the taillight. I have a couple of friends who own Supernova headlights, and at first I went by the assumption that they would be similar in design. However, as I found out later, the Supernova lights are special and in a league of their own as far as how their driving circuitry is designed.
The basic principles behind dynamo powered bicycle illumination are as follows; there are two major components to the system, the dynamo (hub, or tire sidewall friction generator (and other similar varieties)), and the lights. The dynamo generates the electricity in the form of Alternating Current (AC). The standard for all bicycle dynamos (sidewall and hub) is 3 watts at 6 volts. Being that wattage is equal to voltage times amperage (1W=1V x 1A), that gives us 500mA of current. These generators also happen to be constant current generators, so that 500mA stays 500mA (we’re speaking in terms of AC electrical output from the hub, we haven’t gotten to the rectifiers yet) and as the dynamo spins faster (your speed increases as you ride your bike) the voltage increases. It is possible to generate upwards of 100 volts from these little generators if you’re capable of riding your bike at 70 mph (or if you build a DC voltage pump and then you could have a Jacob’s Ladder, or other spark gap entertainment). Speaking in terms of reality, however, you are not usually capable of generating more than 15 or 20 volts, however, incandescent bulbs would tend to burn out or explode at these high voltages, so sometimes you would get lights that would burn out if you went too fast. Because of this, some hubs have or lights have Zener diodes in them as an AC voltage clamp so that it isn’t possible for them to generate more than a certain voltage (and thus avoid making bulbs explode). The only real driving circuitry in the incandescent headlights/taillights was the bulb and the zener voltage clamp. The bulbs were designed to work with the hub, 6v, 3w, or if there was a taillight, 6v, 2.4w for the headlight, and a .6w for the taillight. This is the way it was done for many many years, until the very recent advent of LED’s that are bright enough to compete with incandescent/halogens.
LED’s run on Direct Current (DC), which is fundamentally different than AC in the same sense that the energy in ocean waves are different from the energy in a river. I’m not going to go into any more detail about it, but I’ll just say that Wikipedia and Google are your friends. To turn the AC coming from the hub into DC, we rely on a little network of diodes called a rectifier. This takes the AC wave (which is bipolar, has both negative and positive portions) and through the diode network, flips the negative portions to the positive side, thereby creating a pulsating DC wave. From there, it goes through a smoothing capacitor which smooths out the pulses in the DC.
The Supernova light, as I mentioned earlier, is special. It is an LED headlight that is designed to work with its specific LED taillight. The terminals for the taillight output DC, and the stand-light circuit (a little circuit that charges while you ride, and then discharges while you are not moving to keep the lights illuminated for short time) in the headlight controls the taillight as well. This design makes a lot of sense. All of the driving electronics are kept in one place, it allows for smaller and simpler taillights, and it simplifies and eliminates redundant stand-light and rectification circuitry. However, it isn’t how most lights are being made because it is less compatible. The old standard, I briefly mentioned, if you were running a headlight/taillight combo was to have a 2.4 watt bulb in the headlight and a .6 watt bulb in the taillight. If you disconnect the taillight, you would need to replace the 2.4 watt bulb with a 3 watt, or it would burn out. The Schmidt Edelux headlight is designed to this standard. It is designed to have a taillight that draws .6w AC, so it will work with the vast majority of all taillights being manufactured today, which, despite the fact that any taillight that you would want to own has LED’s in it, is the old standard. You could, in theory, run a taillight made in 1972 with the Edelux if it has a 6v/.6w bulb in it. The silliness of this, if it isn’t apparent already, is that for both the headlight and taillight you need a diode rectifier, and a stand-light circuit, which duplicates parts unnecessarily leading to more potential places of failure and higher costs, not to mention the increase in size and weight to handle these redundant parts.
In conclusion of that rant, Supernova is headed in the right direction, and hopefully other light manufactures will follow suite (but probably not, sadly, because old standards die hard and orphaned products are a pain).
In regard to Alan’s taillight, I learned all of this the hard way. I designed it to work well with the Supernova, because that was the light I had access to and I assumed that they would be similar in design, both being top shelf LED headlights. When we first tested the taillight with the Edelux, it was bright enough, but flashed badly at low speeds, and didn’t share the stand-light function of the headlight. When you power an LED with AC, it will only light during the positive iterations of the waveform, and possibly damage the LED during the negative iterations. Diodes in general don’t respond well to reverse biased voltages, unless they are designed for it, like Zener Diodes are. After I found this out, I added a small bridge rectifier, and a NEC/Tokin 1 Farad capacitor that I pulled out of a dead Busch&Muller taillight I had kicking around for the stand-light power storage. I used around an 80 ohm resistor to limit the amount of current the LED was taking from the headlight. I used a Cree XR-E for this project because of it’s relatively large lumen output with a small current draw. The Red outputs 40 lumens at 350mA and has a max drive current of 700mA.
The final installation gave a little bit of trouble due to a failure on my part to adequately electrically insulate a couple of important areas of the light, but a little bit of last minute heat shrink tubing solved that issue. Being able to see it on the bike and fully illuminated at night was a real treat. The lens on the old Luxor light shapes the beam into a tight bright spot with a relatively low intensity very wide spread, so riding a foot or two behind it isn’t unpleasant because your above where the spot is aimed, but from 20 feet it is dazzlingly bright. It has very good side visibility as well, and the most importantly, the whole shebang looks very sleek and like it was made to go together.