J. Kissel We'd chased our tails for a few days because we were comparing uncalibrated alignment data mode cleaner mirror against calibrated data (see LHO aLOG 5660). We believe the calibration was not present over these past few days (maybe weeks) because the safe.snap was out-of-date. So, I've taken the time to give them a thorough, expert's thrice-over, ensuring all parameters are in the correct configuration, and have captured new safe.snaps. Details below. Items that I've confirmed: - OSEM open light current compensation is in place (OSEMINF OFFSETs and GAINs) - MC1 - M1 from LHO aLOG 4067, - M2 & M3 from LHO aLOG 3409 - MC2 - M1, M2, & M3 from LHO aLOG 3505, - MC3 - M2 & M3 from LHO aLOG 4624. - Sensor calibration is ON (see LLO aLOG 4291 for derivation) - OSEMINF FM5 "to_um" filter is engaged for all stages (see LLO aLOG 4555 for original implementation) - DAMP FM5 "norm*" filter is engaged for M1 stage (see LLO aLOG 5388 for original implementation) - DAMP damping loops are set with up-to-date filters and gains: - L: FM3 [resg], FM5 [normL] G = -1.55 - T: FM1 [0:25,25], FM10 [ellip50], G = -2.0 - V: FM1 [0:20,20], FM10 [ellip50], G = -3.0 - R: FM1 [0:20,20], FM10 [ellip50], G = -0.2 - P: FM1 [0:25,25], FM10 [ellip50], G = -2.0 - Y: FM1 [0:20,20], FM10 [ellip50], G = -1.0 Note, the exact frequencies at which the velocity damping filter (FM1) in T,V,R,P,and Y differ from LLO, but this doesn't really affect the performance -- they're crappy whether they're all at 25 Hz as in LLO or at 25 +/- 5 Hz as they are here (and isn't captured by the BURT snapshot anyways). They will be replaced soon, so for now the discrepancy is not worth fixing. - ISIWIT Path is installed, on and engaged (originally from LHO aLOG 4553, but I couldn't find a log indicating when things were specifically installed for the MC mirrors.) - OPTICALIGN M1 alignment offset drive calibration is installed (from LHO aLOG 4763) - Alignment offsets from a prior good IMC lock are installed (from LHO aLOG 5663) - All stages of inner-most watchdog parameters are set, and the 'dogs are cleared and armed - OSEM DC Threshold set to +/- 30000 [cts] (based off of a ~few thousand count buffer of the full ADC range) - OSEM AC Threshold set to 8000 [cts RMS] (based off of empirical experience of "what it rarely goes over during normal use) - ACT AC Threshold set to 25000 [cts RMS] (based off of empirical experience of "what it rarely goes over during normal use) - All Bandlimiting filters are engaged - Coil Driver BIO states set to 1 (see T1100507) - Low Pass filter OFF for TTOP driver - Low Pass and Acquire filters OFF for TACQ driver - [MC2 Only] Locking Filter paths are as described from a prior good IMC lock (from LHO aLOG 5452) And naturally, for the safe.snap capture, I've turned OFF all requested drive to the DAC, so - OPTICALIGN alignment offsets are turned off (at the OFFSET button level) - DAMP damping loops are turned off (at the OUTPUT button level) - MASTERSWITCH is OFF -------- Though Mark has written a nice utility for capturing safe.snaps (see LHO aLOG 5202), I needed to get this done fast, hadn't used it before, and saw a few things I didn't like (which we've since corrected), so I captured the files "by hand" (which is quickest for me), hence: (1) Turned off all requested DAC drive, i.e. "natural" steps listed above by hand. (2) Captured each suspension's safe.snap, saving over what is in the userapps repository, e.g. controls@opsws0:burtfiles 0$ pwd /opt/rtcds/userapps/release/sus/h1/burtfiles controls@opsws0:burtfiles 0$ burtrb -f /opt/rtcds/lho/h1/target/h1susmc1/h1susmc1epics/autoBurt.req > /opt/rtcds/userapps/release/sus/h1/burtfiles/mc1/h1susmc1_safe.snap (3) Ensured that the link to target directory (which is where the front end *actually* looks during reboot) was in place, e.g. controls@opsws0:burt 0$ pwd /opt/rtcds/lho/h1/target/h1susmc1/h1susmc1epics/burt/ controls@opsws0:burt 0$ ls -l safe.snap lrwxrwxrwx 1 controls controls 63 2013-01-17 12:46 safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1susmc1_safe.snap (4) committed the new safe.snap to the userapps repository,
J. Kissel, C. Vorvick Since I'm not nearly as much of an expert on the HAUX, I asked Cheryl to set up the IMs as she liked them, with appropriately good alignment offsets and captured a new h1susim_safe.snap as well. Two things of note: - The alignment offsets are installed in the TEST filter bank, because the fields built for alignment offsets (i.e. ${IFO}:SUS-${OPTIC}_M1_OFFSET_${DOF}) have not been converted from simple EPICs values to a full filter bank as was done for all of the multi-stage suspensions back in October to take advantage of the ramping features (see T1200478). It would be aesthetically good to change the alignment offset infrastructure to be the same as the rest of the suspensions, but it's not necessary. (though it's scary to have a place to mistakenly add non-working vestigial alignment offsets...) - All watchdog values are set to +/- 100000 [ct] or 100000 [ct RMS], which is well outside of range that would actually protect the suspension. Cheryl says "they've been intentionally set to ridiculous numbers such that we can work in chamber without having to reset the watchdog all the time." Hence once the chamber is closed up an settled, we should set these values back to reasonable, protective, values, and recapture the safe.snap.