Tanks A Lot writes :
However, the main diaphragm is an entirely different beast. It has to be flexible to a fair amount, very sturdy
That is very true, but let us make a first approach to the beast:
The axial forces from the pressure on the membrane are, on the one hand, transmitted to the outer clamping and the main part to the membrane plate.
Now let's assume that the membrane can be slightly bent, so predominantly tensile stresses arise in the membrane. Then, with a vertical tangent, no axial forces are exerted to the membraneplate, and the membrane can be freed here (replaced by a radial force). This gives you Reff.
Within Reff, the axial component of the tensile stress acts on the membrane plate, and outside of Reff, the axial component of the tensile stress acts on the outer membrane clamping .
Now there is still the problem of calculating Reff, which is not easy and too difficult for me.
After all, Reff is greater than Ri and smaller than Ra.
Here is a comparison between the Apex and the Scubapro MK17 evo interesting.
It looks to me like the Scubapro MK 17 has the Reff. of the main membrane and the Reff.
of the environmental membrane with the same value, which also corresponds to the fact that the IP has a constant value relative to ambient pressure.
In this case, it doesn't matter for the IP increase whether there is an air bubble in front of the environmental membrane.
In a deep overcompensated construction (e.g. some Apex), the environmental membrane R_eff is larger than the main membrane R_eff. Here, there is a minimally slowed increase until the bubble is in contact with the environmental membrane. Practically, I would not attribute any significance to this.
However, I would remove the bubble, because when the first stage is pressurized, the environmental membrane will be pushed further forward.