Sheila, Nutsinee
We made a two attempts to inject squeezing today, both times we eventually unlocked the interferometer when the LO servo unlocked. These two locklosses happened at 7:59 UTC and 1:23 UTC today.
We've made some improvements to the automation.
I think looking at LO output and close the shutter when it rails isn't fast enough. We need to look for a pending doom, like a drop in 3MHz signal or error signal. We need a fast channel for this to be able to close beam diverter in time when something goes wrong.
Attached are plots of these two locklosses.
In the first one, the OMC RF3MHz signal drops 4 seconds before the lockloss. We had been checking the LO SERVO fastmon and shutting the beam diverter if it reaches the rails, but this wouldn't have caught either of these LO locklosses. I've changed the SQZ_LO guardian so that it will close the beam diverter if the RF3MHz drops below 20dBm, which would probably have prevented the first of these locklosses. This might not be fast enough to catch locklosses like the second one.
Not relate to this topic. But for the record, Jan 4th was the day we injected squeezing and didn't see anything. We looked into the nlg during the time as we changed 6MHz and 3MHz using Q error signal and turned out we only had nlg of 1.3. This could explain why we didn't see much.