Let’s look at the angles formed by a small parabolic reflector. (A larger C8-sized reflector will produce angles roughly half the size.) Not all my numbers are exact but close enough to get the idea.
Here you can see that the beam angle gets better where the reflector surface is farther from the source.
Further back where the only the emitter’s edge is being reflected, the beam angle looks even better. (The 2* is a guess since I can’t measure the edge of the phosphor.) But lets rotate that reflector 90* so that point is at the red dot and take another look …
Now it shows a beam angle of up to 80*. I’m not sure how much but I think the beam pattern at that point curves with the circumference forming a crescent so it’s not as bad as it looks but still far worse than what any points more forward would create.
Edgewise, the emitter is a long line of points so while it creates a narrow angle in one axis, it has a wide angle in the other. Also, the side of the emitter is in close proximity to the reflector which gives a wide angle.
