Ed, Georgia

Tonight we encountered the same problem as last night - a 0.5Hz instability that rang up in the INCREASE_POWER state. We tried various iterations of gains and notches in INP1 and the DC centering loops, including what was suggested by Sheila and Keita, but still had trouble.

What we did that allowed us to power up (this time)

• Added 60s waits after each power step in INCREASE_POWER, to allow the ASC to converge better at each step (maybe unnecessary?)
• Changed the phasing of the PIT3 dither loop, from 155 to 90 degrees (probably not optimal)

For now both of these changes are in the guardian/accepted in SDF.

The power up whole story

I added stricter convergence checkers in INCREASE_POWER but they weren't strict enough and we lost lock again. I've left these in the guardian anyway.

I noticed that when we engage the soft loops the POP RF18 buildups get worse (though the recycling gain gets better). I tried adding an offset to the PRC2_P loop (still at 2W input power), saw RF18 and the power recycling gain increasing and the 0.5 Hz instability ringing up in pitch and INP1 Yaw, strange.

I opened the PIT3 ADS loop (pointing of PRM) and aligned it by hand for better buildups.

We tried powering up again, with 60s waits at each 5 W power step, and survived just fine. At this point (sitting at 20W) I tried closing the PIT3 loop again, and as it, and the rest of the ASC, converged we saw the 0.5 Hz ring up and we lost lock. I very roughly re-phased the PIT3 loop so that the error signal was zero for the better alignment.

The next time we survived the power up.

This PIT3 phasing does not look good full power: there is a clear line in DARM at the dither frequency (19.653 Hz), and there is some lumpiness at low frequency that looks like bad A2L. I think we haven't actually solved the problem, but since we had a superevent (just locked in time, whew!) I don't want to take us out of observing to play with the phasing.

Intensity noise coupling

This lock has a terrible breathing noise above 2 kHz. This is intensity noise (see attached plot and coherence between DARM and the ISS second loop PDs) and is surprising since it should be squashed by the ISS second loop. Maybe we need to take a look at the open loop transfer function and adjust the gain?