Hi sendithard,
It's not about doubling the worst Vs. recording a surface deviation from basic.
The profile measurement always doubles the worst deviation.
It's about the different ways the surfaces are controlled by the two profile tolerances.
I think I assume correctly that the measured surfaces are the ones colored in purple.
In the AB Datum Reference Frame, the tolerance zone is floating in the longitudinal direction, because the degree of freedom in that direction is unconstrained. Therefore the measurement doesn't consider where the smaller surface is relative to datum C (datum C is not even recognized by this measurement).
In the measurement relative to ABC all 6 degrees of freedom are constrained, therefore the tolerance zone is fixed relative to all 3 datums, and any variation of the smaller surface in the longitudinal direction is recorded. Therefore the larger measured value.