Just some continuation, as maybe someone later is interested in this stuff, and so can find it in a search machine...
Finally the el-cheapo Oximeter-Sensor arrived from china.
It was a tough job to open up the cast sub-d plug without damaging anything...(needed half an hour just for that)
Finally this is how the internal PCB looks like (see attached pics). As you can see, it has a nice little 8-pin soic on it.
Had a short look, and the only possible ICs for that would be a DS2433 (and it's derivatives) or a DS2502.
One is a 4KBit EEPROM, the other a 1Kbit OTP EPROM.
So my guess would be, they did go for the OTP EPROM.
Finally the question has been answered:
Of the unused Pins on the DB9 connector Pin 4 is GND and Pin 8 is Data for the 1-wire.
But what is really strange is, that the SOIC is wired 180° around.
Meaning the data pin should be on pin3 of the SOIC and the GND on Pin4, but actually they are on Pin 7 and 8???
So either these are not "original" 1-wire chips. Or they used a batch which has been wrongly wired.
If these are not "originals" then they could anyways be of any type...
We will see this if I read it out.
Edit:
Ah, now, when I look at the strange setup with the diode and the cap, I just realized, how they did it. They just took a small uC to emulate a 1-wire-memory-device.
And the diode+cap is needed for the generation of the VCC of the IC.
I at least hope it is like that. And not that Nellcor is using some proprietary 1-wire stuff implemented in a uC for their sensors. Without knowing the protocol, one would then really be pi**ed...
But what really puzzles me is C2. It is a capacitor directly connected between Data and GND. Does this make any sense in a 1-wire device? Maybe just for filtering...