SU(1,1) interferometry, proposed in a classic 1986 paper by Yurke, McCall, and Klauder [Phys. Rev. A 33, 4033 (1986)], involves squeezing, displacing, and then unsqueezing two bosonic modes. It has, over the past decade, been implemented in a variety of experiments. Here I take SU(1,1) interferometry apart, to see how and why it ticks. SU(1,1) interferometry arises naturally as the two-mode version of active-squeezing-enhanced, back-action-evading measurements aimed at detecting the phase-space displacement of a harmonic oscillator subjected to a classical force. Truncating an SU(1,1) interferometer, by omitting the second two-mode squeezer, leaves a prototype that uses the entanglement of two-mode squeezing to detect and characterize a disturbance on one of the two modes from measurement statistics gathered from both modes.