Hi,

Now I get know how your circuit works. Correct me if I am wrong. You will use a fixed (ofcourse, variable gain will give bit of vesatality) gain photo detector amplifier which feeds the ADC. By using the proposed "span" circuit you will adjust for ADC full scale for "blank" and measure the actual for "sample" and rest of the calculations are straight forward. This has many advantages. Clever implementation.

Only negative side is, you may have to waste so many IO pins. Can't you use any serial DAC?

Off note, how much of variation do you expect between lamps?