To answer the post's final question: Yes. It is quite common to design instrumentation to cancel systematic uncertainties or nonlinearities to leading order. F

To answer the post's final question: Yes. It is quite common to design instrumentation to cancel systematic uncertainties or nonlinearities to leading order. Fancier arrangements go to higher orders. In the particular case of the differential capacitor, that form of differential measurement is particularly common -- if one non-linear system can paired with a symmetric partner with equal and opposite nonlinearity, the leading-order non-linearities are suppressed (to the extent that the matched pair are actually matched). An easy-to-understand example of the compensation of linear effects is the temperature-compensated pendulum clock: https://en.wikipedia.org/wiki/Gridiron_pendulum https://en.wikipedia.org/wiki/Gridiron_pendulum