The ITM ring heaters were switched off on 2016/11/06 13:55 UTC (with two accidental short streches reangaging them 17:27-17:38, and 18:48-18:53 UTC).

The ETM ring heaters were set to 0.5W per segment (1W total) [from 0.75W/segement - 1.5W total] on 2016/11/06 19:14 UTC

Additionally, we set the in-lock CO2 TCS settings to 0.268W (X-arm) and 0.068 (Y-arm).

This corresponds to the setting we had during the 5h long nice lock at 25W with hardly any couplings at higher frequencies.

Operationally, the main difference were:

 - Our current PI settings drive mode 27 unstable - just shutting off the feed-back prevents runaway.

 - the SRC1 ASC loop goes unstable about 1h into the lock. THis behaviour is familiar for past operation. For our noise test we can simply shut SRC1 down temporarily.

Attached are plots of calibrated DARM, as well as DCPD cross power spectra (not corrected fro the closed-loop gain).

Note that the TCS revert did NOT result in lower noise couplings above 200Hz. (In fact, it seems somewhat worse than the November settings - but that might just be due to more duning in the November state).

Bottom line: It is still a mystery to me why the jitter noise above 200Hz was 3x to 5x better in May.