Let's Make Robots!

UV Fluorescence marking - initial concept fail

I've done a bit of initial feasability testing for this project. Last week I used the four 50mcd UV LEDs and two 630nm detection LEDs. The signal generated from the op-amp was around 0.15V. Background noise was measured at about 0.05V. 

I've since received the 2000mcd UV LEDs, and wow, are they bright! They also don't blind the detector LEDs during direct exposure. The less powerful 50mcd LEDs would.

A picture of one of the longer distance test setups:

This setup yeilded a signal of ~0.3V (I expected it to be about double what the dual LED detector setup got). Not incredibly powerful but useable. The problem came when the dye was put on the tiles the robots will be rolling around on. For some reason the red tiles give off almost as much detectable light as the dye sample. Maybe a few hundreths of a volt less. Just too much background noise.

My next two options in mind are:

  • Use a filter that only lets the red spectrum pass through.
  • Use a photocell and green fluorescent dye for detection.
  • Retry initial setup with reversed green LEDs and green UV dye.

As far as the first option goes, I've yet to investigate filters but I remember that certain lens filters for cameras weren't too expensive. I might try and find an even more frugal method like red cellophane because I'm cheap. I'm not confident either one will be a successful fix.

The second option is "iffy" at best. Previously I tried using a photocell to pick up the glow from the red dye but it didn't react much. CdS cells have a sensitivity closer to the human eye. The peak detection wavelength is in the green spectrum and it tapers off in the red and blue wavelenghts. Not really a surprise that the weak red fluorescence didn't wiggle the needle. It looks like the green dye glows much more vigorously than the red dye. That would be beneficial. I still need to make a setup to see how much the UV reflecting of the paving tiles drop the photocell's resistance by itself. It will probably be substantial, unfortunately.

The third option has a problem that in that it takes more light to have green LEDs detect the glow. Only the (supposedly) brighter glow of the green dye could possibly make this option work.

The project isn't dead but I'm disappointed all my great ideas were only interesting at best so far LOL I knew detecting the soft glow of fluorescence would be difficult but perhaps I underestimated the challenge given my low-tech/low-cost approach.