I was about ready to make the same experiment, but it seems you already proved my theory. Just curious, but how far is the surface of the lens from diode? My suspicion is the larger lens area is grabbing a much bigger chunk of the emitter area, and the results are obvious downstream with more energy hitting the cardboard. Exit pupil is going to be bigger, but it makes up for it at the focal point. If I recall my optical design stuff (been a lot of years) it's possible to design a lens with a wide field objective *and* small exit window, but requires a fairly complex intermediate grouping in the middle. Telescope eye lenses can do this, but aren't cheap. Nice tests.....shows the limits of our typical optics.