[Alastair, Aidan]

Today we added the chiller loop to the laser servo on Y-arm.  It takes an error signal from the PZT output voltage minus 35V, to set the PZT to the middle of it's range.  From the step function measurement the time constant for a temperature change request to actually change the laser temperature is around 110s.  Gain was taken from previous measurements of power as a function of pzt voltage and chiller temperature https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=11625.

We initially added a servo with unity gain at 0.001Hz, which should have given a reasonable phase margin, however we found that it was oscillating (very slowly), so we altered the unity gain down to 0.0003Hz and it now seems, from a few hours of operation, to be stable.

Attached is a graph of the 3hrs of laser power with the laser unlocked.  The pk-pk is roughly 0.17W.  Also attached is a chart of the laser with the PZT and chiller locked, showing pk-pk of around 0.02W, which in the full data plot looks to be the noise floor.  This plot starts just after the chiller servo has been engaged and shows it bringing the PZT voltage down from 55V, and actuating on the chiller temperature.