This is a viscometer system. The coefficient of viscosity/temperature of the materials is such that a 1 deg change of temp is 5% of reading. As we are after the very best reproducibility, pushing the temp spec to +/- 0.03 C accuracy helps enormously. State of the art is I believe 0.001 C in NPL/NBS. Our saving grace is that we are in a tightly defined spot temps, but our system is pretty stable away from the spot points.

Steve