08/22/2018

Since I am waiting for AS72653 to show up at the usual retail sites so I can assemble some more of the "production" version of the spectrometer I decided to take a look at a variant design. I noticed that both the original and new AMS AS7265X evaluation boards arrange the three 6-channel sensors in a trifoil pattern. This seems like a good idea to both equalize the distance to the target object and light source and to minimize this distance. So this is what I tried to do. I redesigned the AS7265X spectrometer to place the sensors in a close trifoil pattern.

The board is smaller (0.70 x 0.70 inch) than the original production version (0.71 x 0.76 inch) and is actually quite a bit smaller in area because of the rounded edge. I replaced the switch to toggle between I2C and UART modes with a solder jumper, since this is cheaper and in my experience one uses one or the other and doesn't jump around from one to the other. I also got rid of the IR led footprint. This is not because the IR led is not useful, it is. It's just that at ~$5 the cost of the IR led is just too much to fit within the "inexpensive" spirit of this project.

I kept the interface pins along a single edge and got rid of the 0.1-inch mounting holes which I never used. I found that soldering 22-gauge hook-up wire to the board in a heat-shrunk bundle provides a nice stiff cable with which to hold the spectrometer. Any additional mounting can be done by using a 3D printed frame around the edge or something similar.

The distance to the three sensor ports from the center of the board is a constant 0.18 inch (~5 mm) and the broad-band led is right in the center. So the target will be well illuminated and each sensor will "see" the same reflected light. This is as close to ideal geometry as I think it is possible to get with three separate 6-channel sensors and I hope that I will obtain even better results with this set up.

I should get the boards back from OSH Park in a couple of weeks for testing.