I wonder if some of the beam pattern issues are because the light is supposed to automatically adjust its beam pattern based on speed. Presumably it gets some info from the current input (sidewall or hub dynamo), and when rpms increase the beam pattern lengthens versus the lower speed broader pattern. This is all academic for me since I don't have my setup yet to install and verify any of this myself. Since most dynamo tests are performed from a static unit (wheel turning to generate current), I wonder if some of the ability of the light is not being called into play. Perhaps the crappy beam shot is the lower speed versus higher speed. I'm gonna go out on a limb and say it's got to be better than the 120 lumen battery powered Cygolite I've been using, and if it's even marginally better than that and I never have to worry about whether I remembered to charge my batteries, well I'm coming out ahead then.