I've measured 100v open-circuit on my SON28.. but it was at 50mph. There's nothing quite as fun as riding down a steep hill at 50mph and trying to read the multi-meter rubber-banded to your handlebars!

I'm going to guess that someone in Germany knows what fails first when trying to maximize output power at high speeds. Maybe the copper wire gets warm enough to produce shorts in the enamel insulation? There is probably some hysteretic core loss at high speed that would contribute to heating the wire.

As for open-circuit voltage... the headlight circuit design would determine how sensitive it is to high voltage. I took apart a dead B&M Lumotec Eyc for fun. It uses a little momentary switch to control the function, so some internal circuitry is always powered up. It has a bidirectional protection zener at the dynamo input. It appears to be a 100V transorb. That would make you think the circuitry is rated to handle 100V, but the transorb is located at the input to the full wave bridge rectifier, and at the output of the bridge rectifier is a 680uF, 10V cap. I'm not sure what the transorb is protecting if that cap is only rated for 10V. I can't decide whether there is enough current draw by the Eyc to keep the dynamo voltage pulled down to a low, safe level, ... or whether B&M assumes that no one is going very fast.

I think the question about whether open-circuit voltage matters will end up being the standard engineering answer.... "it depends".

Steve in Peoria