my bad for being a bit myopic. yeah the classic version was not what i was focussing on...it's funny how we read what we want to read
i can see how a fourth single cell with deco would have some advantages. The deco is closer to real time po2 and the it's a totally isolated system in terms of current. It has the vulnerability of being one cell, but it's not being used to drive the solenoid so it does not require the redundancy of three cells as is required for voting logic for a e-set point controller. since you can verify the reading with a flush and test for current limitation you can tell if the single cell is going a-wall... or so the philosophy of this paradigm seems to go.
I admit that a single cell sharing one of the three seems kind of silly. Leads me to wonder, if you are going to accept the concept of sharing a single cell, why not share all three.
hopefully I got you point.
g