Remark: Why this is not a scheme change

Changing $\epsilon$, the ramp, or the preparation changes the source ${\sigma }_{\epsilon }h$. It is not an admissible finite renormalization of a fixed source experiment. For overlapping source and observation supports, a residual pairing can still contain local calibration dependence; the late detector in Theorem thm:residual is the separate construction that removes all such contacts.