Following on from Hugh's alog entry (https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=22883) about the reference
cavity temperature and HEPI hitting its software limit ... Attached are two plots of the reference cavity temperature:
an in the loop measurement (H1:PSL-FSS_DINCO_REFCAV_TEMP) and out of loop measurement (H1:PSL-FSS_DINCO_REFCAV_TEMP_OOL).

The coefficient of thermal expansion for Corning 7940 is 0.52E-6 / degC.  For the ~0.01 degC change this corresponds
to a length change in the reference cavity of 5.2 nm, or a dL/L of 2.6E-8.  Assuming the equivalent dL/L for HEPI is
6.3E-8 (250E-6/4000) then the temperature change doesn't fully explain why HEPI hit its limit.

However temperature regulation of the reference cavity wouldn't hurt, particularly if done at a point higher than room
temperature.  The hardware is already there, modulo a small change to bring the heater front end model to reflect the
installed cabling - a post observing run task.