The values in the table are volts measured at the load. Probably Vrms, I assume.
The left column labeled "frequency" is uncertain, but I'm assuming it's the frequency of the SON output. IIRC, the SON has 26 poles, so at 1 rev/second, the output would be 26Hz. Running through the calculations, with a nominal wheel diameter of 27 inches, that should mean that 25Hz yields 4.63mph, 30Hz yields 5.56mph, etc, up to 60Hz yielding 11.1mph. Pretty slow stuff for most of us.
Calculating the power produced at each data point shows that over 7 watts was being produced with the 35 ohm load at 11.1mph. The data hints that 35 ohms is not the load that will yield the max output power, and that a larger impedance load would be better. I wonder how much power you could get at 20mph??
Now this is really getting useful, and fascinating. I agree totally with your calculations of wheel size and resulting speeds, since all the resulting data confirms the published specs for the "standard" loads of one 6V/3W bulb, or two in series, at various speeds.
I put all this data into a spreadsheet so I could make it calculate the amperage and wattage, and so I could create graphs. Without further ado:
The voltage is the predictable one, scaling up with speed and/or load, with low loads tapering off almost level at higher speeds. The amperage is very similar, scaling up slower, with the smaller loads tapering off almost level at higher speeds. The interesting thing here is the really low amperage from high loads at low speeds. You're right that this is not truly a constant-current device, but the reason it's called that is because within certain ranges of speed and load (normal cycling speeds with the 12ohm load it's designed for), the amperage is very close to constant. Also, it seems the maximum it can ever produce is around 570mA, which is consistent with published specs, so maybe it should be called a "maximum-amperage device".
But the wattage is the really fascinating one. Again, the low loads are somewhat predictable, we're approaching the saturation point of the generator at 11mph with them, and I'm sure that they would truly level off at higher speeds. The high loads, though, show a marked drop in wattage at low speeds. Up to about 8mph, the 35ohm load produces less wattage than the others, and at 4.6mph, the 18ohm and 20ohm load are pretty close to the maximum wattage. So we've found the curve, or at least part of it, for maximum wattage at a given speed.
Extrapolating off the ends of our data ranges, I imagine that higher loads would show an even steeper wattage curve. For example, 50ohms at 5mph would produce somewhere around 12V, 90mA, and 1W, but at 20mph might produce 25V, 570mA, and 15W. The output from lower loads becomes flat at 11mph, but higher loads obviously have a higher saturation point, with the 35ohm load probably not flattening until over 20mph.
Now I really want to get one of these and start experimenting. I could use household incandescent light bulbs as my load since they can take AC or DC up to 120V. A 300W bulb is a 48 ohm load, which might be a good starting point to see if I can go fast enough (it's easy to hit 40mph on a good downhill) to find where that load starts saturating the generator, and the output starts flattening. I could work back from there using automotive bulbs with lower impedances to create a sort of chart with what load gives the max wattage at various speeds. The neat thing is that the graphs show a very consistent maximum wattage of about 0.7 the speed in mph. So if you could vary the load based on the speed, it seems that would continue to scale up power linearly. This makes sense, because the generator should be able to extract a constant percentage of the power available (determined by the speed). So that 40mph descent, if using the right load, could produce 30W. (From a generator rated for 3W!!) Then I'd have to design a regulator circuit to take advantage of that. If I could get them mass-produced on a little circuit board like TaskLED does, I bet I could sell them like hotcakes.
It might be difficult to find a regulator that can run at input voltages between 12v and 100v. This isn't a typical power source. Things are improving, though. A number of switchers designed for automotive applications are tolerant of 80v or slightly more, so you may be safe in most cases. Just don't go down hills at 50mph.
Yeah, I've been looking and looking with no luck. I might have to use a "regular" regulator and cap the input voltage to protect it. Seeing the power available is well over what I'd thought, I'd feel fine about "throwing away" the extra available at high speeds.
Sorry for running away with the thread like this. Baker and ktronic, if you're still reading, you can make use of this data too. My previous estimates of when your LED will come on and reach full brightness should be lowered by about 2mph, it seems, and that would match ktronic's experience with his LuxV. Again, the Shimano hubs you have work on the same principles and are designed for the same power production at the same speeds as the Schmidt that this data was taken from, so your actual numbers might be slightly different, but should be close enough as to be the same.
When contemplating this kind of trickery, just be aware that Shimano, and SRAM, have just announced their new models of Hub Dynamo, and apparently these are high efficiency DC output designs (similar to the Lightspin, and Lumo S6/12 models). You need a 'traditional' AC output model to be able to do this AFAIK, so look for the current/ older models from these manufacturers.
Thanks for the links! Of course they just say 6V/3W like everyone else, but I'll try to dig up some more technical data on them. It does certainly seem that with the move to LED lighting becoming more mainstream, it would make sense for newer dynamos to output DC directly so it can more easily be used by LED lights.
Alex
