But it never goes to zero

Steve and I first crossed paths in 1984 at the Naval Postgraduate School (NPS). He was building his career as a professor of physics; I was finishing my doctoral dissertation through Penn State while a visiting scientist at NPS. A fortunate intersection for me. Since then, I have been privileged to have learned from his facility and fascination with tools: tools in physics, tools in engineering, and tools in the machine shop. This is one story of many: eye opening for me, business as usual for Steve. After a lengthy conversation about demodulator noise in fiber-optic sensors, I asked about more fundamental issues. In 10 min, he calculated a lower limit to the self-noise based on equipartition, the fluctuation–dissipation theorem, and an approximation to the dynamics of the sense element. While this lower limit was well known in some communities—and, in fact, all the way back to the mirrored galvanometer—publication of the limit caused many developers of new-technology sensors to re-think designs and modify claims of extremely low self-noise. The self-motion of a sensor mechanism does not go to zero with time. Neither does the influence of the master of his craft on the apprentice.