The paddles where the components go can be stiff because they're quite small--a photodiode (5 mm or 14 mm), an SC70 op amp, a SOT23 bootstrap FET, two resistors and a couple of bypass caps. It's mostly the survival of the 'cable' parts that I'm concerned about. If necessary I can probably use wires and leave the FR4 to do the gross mechanical locating.
The whole thing is going to be potted in black silicone rubber (with the PDs visible, of course) because I'm looking at a path loss of > 200 dB electrical--with an amp of LED drive current, we expect to get tens of picoamps of signal photocurrent, and less with a dark-skinned mom.
Melanin absorption causes overestimation of SpO_2, by as much as 8% @ 2 sigma at low SpO_2 values, right where it has the worst effects. The effect mostly disappears at high SpO_2.
We're going to be using a three-wavelength system to avoid that problem, but even without that systematic shift, the absorption still hits the SNR pretty badly.
Cheers
Phil Hobbs