Based on the “design considerations” PDF at https://cdn.sparkfun.com/assets/8/5/f/0 … ations.pdf
it looks like each of the three separate sensors has a field of view of ±20.5 degrees. Based on the photo of the board with the cm length scales, it appears each sensor aperture is 13 mm distant from the other two. It looks like the chips could be physically about twice as close together as they are. Why wasn’t the PCB designed to put them closer together?
I presume you would normally want the three wavelength bands to be measured over the same area of the sample. However, for any small area sample where the detectors must be close to the surface, the field of view of the three sensors will only partially overlap. To be specific: if the surface is at or closer than 18 mm, there will be no overlap at all between the fields of view, because 13 mm / 2*sin(20.5) = 18.6 mm
I note that the input aperture is 0.75 mm and it is possible to get 0.75 mm diameter fiber optics (PMMA - plastic) so you could make each of the three sensors fiber-coupled that way, and bring the fibers together for a small effective sensing spot size. Maybe 3D-print a holder to fix the fibers in place on the sensors. This seems like a reasonable enough approach that someone may already have done so… have they?
One other note: the product description says “The Triad […] can detect the light from 410nm (UV) to 940nm (IR).” In grade school they teach the color names as Roy G. Biv or red, orange, yellow, green, blue, indigo, violet. The visual color red ranges from 625-740 nm so it is certainly correct to label 980 nm as IR, but violet goes from 380 to 435 nm so it is incorrect to call 410 nm “UV”. The wikipedia entry on “ultraviolet” shows a variety of wavelength ranges, but I have never seen anything longer than 400 nm called UV (at least, among people who were not trying to sell something.)