Another event similar to the one of Feb 10th occurred to the cooling system of the Slave Laser tonight. I was called by A. Magazzu' because the temperature of crystal 1 of the SL increased above the threshold of 25.5C. I came on site to see if I could patch the situation by moving the tubes attached to the SL head, closing and opening the valves and so on. While doing so, the temperature of the crystal went too high and triggered the safe state of the SL. I restored the correct operation of the SL and then monitored again the temperature of the crystals (see first plot). Since the temperature was rising and the corrections seemed to go towards saturations, I went back to the laser lab and physically unplugged the tubes from the mains under the floor, made some liquid to go out and then plugged back the tubes. The operation took a couple of minutes and was effective in (temporarily!) patching the cooling action.

Thanks to the channels asked by F. Cleva and put online by G. Ballardin (48388), it is possible to infer the effectiveness of the cooling circuit. In the first plot the PSL_SL_*_W are the channels monitoring the actuations of the TEC. The corrections for the pump diodes stay constant over the whole period, this is an indication that the cooling circuit of the diodes is ok. By the way, when the SL set to safe state, the corrections went to 10 a.u.: is this the zero-correction value? On the contrary, the corrections sent to the TEC of the crystals are saturated to 0 a.u., so this is correctly identified as maximum correction value, and this means - as aleady pointed out by Frederic - that the cooling circuit of the SL head is in bad need of some cleaning action to recover its reliability.

Looking at the channels of the actuations, problems of temperature can be spotted well in advance. In the second plot the trend of the correction is clearly visible well before the temperature of the crystal started to increase.

Finally, the powers of the PSL seem to have recovered the usual values, the ITF could be locked back to science mode (see third plot).

As soon as it can be fitted in the run schedule, a cleaning of the cooling circuit of the SL head is needed.