I've updated NUC34's PI monitor with lockloss references, see CR screenshot with red highlighting the danger areas. When the upper lump of the live black trace reaches the reference, this is consistently where PI29 rung up locklosses, in the first 1-3 hours of lock. The bottom right corner monitors for PI 29; it is bad if any peak appears, though we have survived brief appearances in the past. 

It's handy to look for the 80 kHz PI 29 on an aliased-down version on the slower DQ channel (pops up at ~14.76 kHz, 68165), b/c we can't play back the 500kHz fast channel. The black *live* trace thermalizes from lower to high frequencies.

Brina and I are now starting to look into the recent PI29 locklosses, starting with 06/02/23 20:13:27 UTC (noted by Brina in 70153). This is after the change from 76W to 75W input power (5/31, 70042), which followed the ETMX ring heater change to 1.2W (5/24, 69871). To keep damping this PI, we should check the signal bandpass filter for this mode -- the DCPD's see the alias'ed ring up 10 Hz lower than where it "usually" is (examples here), which is strange and would explain why its damping didn't work, if the mode was out of band. I don't think it's just that our increased damping strength (69800) pushing the mode out of band (we've seen weak damping move the freq before by a couple Hz, 68165). Watching the lockloss, I don't see the mode frequency move down by 10 Hz, I only see it pop up at that frequency already. Will investigate more this week.