Hi, I used to work servicing pulse oximeters, and I can tell that although the principle of operation is simple, the actual practical application is not.
You can read about the principle of operation
here, as you can see you need both the infrared and red light source, since the SPO2 is a function of the difference between the emitted and received light of both sources (you need the 2 different wavelengths).
What I can tell you is that, after digitalizing this light signal, there is a lot of signal processing to do if want to get a usable measurement, the algorithms are quite complicated and a play a big part in the quality of the waveform, the oximeters I have worked with were always constructed with a main board, and a separate oximetry board, (usually manufactured by a third party, I’ve seen mostly Respironics and Nonin and Nellcor boards), the oxymetry board would do all the measurements and signal processing and feed the main board (based around a 80C31) the measured values and waveform, so the main board would just be a display driver and an alarm and power manager.
I have only serviced low-cost oxymeters, and some of the boards (specially the cheaper ones with no brand) were barely unusable, the wave forms were always jittered and the SPO2 value would not be stable nor accurate (we used an index2 simulator by Fluke). There’s a lot of research involved in those algorithms.
Sorry about the long (and not very helpful) post, if I wore you I would try to get one of those boards or maybe their documentation, its not a trivial task what you are trying to do but I give you my best wishes that you succeed. Please keep us posted on your progress.