camilla.compton@LIGO.ORG - posted 12:59, Thursday 07 December 2023 - last comment - 11:12, Thursday 21 December 2023(74658)
Turning off all HWS for DARM Comb test
TJ, Ansel, Dave, Camilla
After finding that 1Hz and 5Hz DARM combs were caused by HWS camera sync frequency in 74617, we've turned off all HWS at 21:00UTC so we can check that nothing else from HWS is effecting DARM.
Stopped all HWS code (ETMY was already off 73915). Tuned off all cameras and RCX CLinks. Turned off the ITM SLEDs, ETMX laser will still be on but was tested in 69648 not to effect DARM.
As long as the computers aren't restarted I think that the H1:TCS-E{OPTIC}_HWS_ channels will remain reading their last values, if not Dave might need to "green-up" the CDS overview.
I temporarily took out the DIAG_MAIN notifications for all HWS. We will have to remember to put them back in though when we are done turning them on and off.
Unfortunately the 7 Hz comb is still present in DARM (figure 1) and in the magnetometer channel where we were tracking it (figure 2).
Extra: just to check that this really is a 7 Hz comb that started after the camera sync frequencies were moved to 7 Hz on Dec 5 (74617), and not something pre-existing, I re-plotted the magnetometer data for the time period when both were set to 5 Hz. This time period avoids any obscuring effects of the 1 Hz comb in magnetometer data. (The 1 Hz comb is much weaker than 7 Hz in DARM so it shouldn't really be an issue, but in the magnetometer data both are clear and there's a little more ambiguity there-- and anyway it's easy to check.) Indeed, no sign of 7 Hz during that hour except at frequencies that are also 5 Hz multiples (figure 3).
I put modified HWS() notifications back into DIAG_MAIN now Dave has a running script that turns the HWS camera/code off within 60 seconds of ISC_LOCK reaching state 580 and on within 60s of IFO being DOWN 74951. DIAG_MAIN may give the warning 'HWS code stopped for ITMX / HWS code stopped for ITMY' as soon as we loose lock for up to 60s, if this is annoying, we can make DIAG_MAIN cleverer. svn link