Hey guys, I was playing around with my Shimano DH-3N70 dynohub in the lab the other day, and thought I would share my results for anyone else interested. I wasn't satisfied with the information gain from the "3W 6V" rating on the hub.
I have little (read: no) experience with motors/generators, but a fair knowledge of electrical theory. Generators act like current sources rather than voltage sources, and produce AC current, so there is more to the story than "3W 6V."
The following are rough estimations. I have access to an accurate oscilloscope, but the tests were by no means identical. I turned the wheel by hand to what I thought was a speed producing full power under each of the different loads. Here's what I've got.
When hooked up to a lumotec light (no standlight) alone, the dynohub produces roughly a 8.5Vpp waveform. (RMS of about 6V)
When a 10 Ohm resistor is shunted across the light, voltage drops to about 3.5Vpp, and light output is considerably less (not enough to light up the road, but just enough to be seen)
Using this information, and assuming the impedance of the dynohub is negligible at the operating frequency, the impedance of the lumotec light is roughly 13 ohms, which coincides with the 3W rating.
With an open load, the voltage waveform takes on a very different shape (reminds me of a, tangent graph), with voltage going all the way up to 80Vpp. Clearly this is behaving as a current source.
Basically, the dynohub was pretty consistent in providing about 750mA of RMS current. (all numbers are quite rough, remember). As expected, the waveform takes the shape of a well-behaved sine curve under normal loads at operating speeds, with frequency increasing with wheel speed. Peak current is slightly more than 1 A, and RMS around 750mA. I have a feeling there is a voltage clamp within the light itself to keep the max voltage around 6-7V.
I hope this is useful to someone. Sorry for the inaccuracy of the whole thing. I just got bored at lunch and decided to play around. To make things worse, I had to get these numbers from memory, since I can't seem to find the paper I wrote everything down on. Has anyone else looked at the dynohub like this and can confirm these results?