What fatigue (S-N) curve are you using? Is S labeled stress amplitude or stress range?
In CAESAR II we typically run an expansion stress range calculation. You can throw pressure in there as well.
Amplitude is zero-to-peak, range is peak-to-peak. So if you are evaluating range, use twice the amplitude value.
You seem to question the limits of your range - whether they are both on the same side of "ambient" or on different sides of ambient. Those limits define the boundaries of the range calculation and it doesn't really matter if they cross ambient conditions as long as you evaluate the complete range.
For years, CAESAR II recommended range calculations between each operating state and the ambient (or installation) condition and it was up to the designer to build the range calculation between operating states if one of those states is on the "other side" of ambient. The current version of CAESAR II (by default) now recommends all ranges between operating cases as well.
_________________________
Dave Diehl