Δαφνη Ευτυχια Αρθουρος

I can't afford to order parts yet, but I figured I should work out enough of the design for the infrared flash I want to build, to have some idea what parts I need to save up for. I know that infrared LEDs are available with different angles of coverage for the cone of light they project, and I can easily enough calculate the angle-of-view for lenses of various focal lengths ... but I'm missing a crucial clue required for this geometry problem.

Do IR LEDs project a sharply-defined cone of even illumination; or do they appear to get continuously brighter as you get closer to seeing them head-on, with the viewing angle on the data sheet just indicating the range in which they're brighter than some industry-standard arbitrary threshold? (If you graph intensity versus viewing-angle, does the plot look more like the blue curve or the red curve in the figure to the right?) When you're using an IR photoreceptor simply as a switch, it doesn't matter -- either the coverage angle just tells you the angle over which your detector is guaranteed to get at least a certain amount of the emitter's power, or if you're trying to detect the orientation of the emitter you calibrate the receptor to trigger at the threshold you observe at the angle you want to declare close enough. But for photography, it's going to affect how hard it is to avoid "hot spots" in my photos, overexposed areas, uneven lighting.

Is the answer the same as for visible-light LEDs? If so, I can find out experimentally with LEDs I have at hand, easily enough. I'm betting that coverage angle depends at least partly on the shape of the lens -- domed vs. flat -- but does the fuzzy-vs.-sharp distinction also depend on the package? Or is the answer an extremely convenient "it's always like the blue curve" or "it's always like the red curve"?

I've been thinking of mounting the IR LED array on a flexible (or possibly hinged) surface, so I can change the curvature to change the coverage angle of the flash. So that when I'm using a 200mm lens, I'm not wasting energy lighting up the whole area that a 28mm lens would see, and can therefore get more distance (or use a smaller aperture) with the same number of milliWatts. I'll need to be careful not to create hot spots if I do that. And even if I don't make it adjustable, of the LEDs themselves make a hot spot in the center, I'll need to play games with the angles of different LEDs in the array to even that out. Unless I sacrifice some percentage of the output and stick a diffuser in front of them ... or is a diffuser with a Fresnel lens in front of that the proper way to go about this regardless?