L. Dartez I reduced the amplitudes of the SUS-ETMX calibration lines between 15-17.6Hz. These line amplitudes hadn't yet been adjusted for the reduced noise level at low frequencies (LHO:71200). As such, they were significantly higher than necessary to maintain an uncertainty of 0.5%. N.B. The threshold itself is rather arbitrary. This is discussed further in LHO:69555. Moreover, there was evidence that the high SNR lines were causing excess noise (LHO:71149) and repeated on/off tests with and without the calibration lines active showed an improvement when the lines were off (LHO:71296). The new values, as compared to their old settings set in LHO:69555, are shown below. Channel Old Gain New Gain Reduction Factor ================= ========= =========== =========== H1:SUS-ETMX_L1_CAL_LINE_CLKGAIN 35 6.6 5.3 H1:SUS-ETMX_L2_CAL_LINE_CLKGAIN 50 9 5.55 H1:SUS-ETMX_L3_CAL_LINE_CLKGAIN 0.3 0.085 3.529 The third attachment shows a short trend of the uncertainty calculations after changing the line amplitudes. They are all hovering between 0.4% and 0.5%. It's worth noting that the IFO was thermalizing during this activity; we had just relocked when I got started. I didn't have time to visit reducing the amplitude of the PCALY line at 17.1Hz before the end of the allotted commissioning time.
While I was adjusting the line heights, Jeff let me know about the following tip:
Stepping the calibration height for one of the ETM{X,Y} SUS lines consists of changing all of the following channels such that they're all reflecting the same amplitude gain value.
Using the TST stage as a reference:
H1:SUS-ETMX_L3_CAL_LINE_CLKGAIN
H1:SUS-ETMX_L3_CAL_LINE_SINGAIN
H1:SUS-ETMX_L3_CAL_LINE_COSGAIN
H1:CAL-CS_TDEP_SUS_LINE3_COMPARISON_OSC_CLKGAIN
H1:CAL-CS_TDEP_SUS_LINE3_COMPARISON_OSC_SINGAIN
H1:CAL-CS_TDEP_SUS_LINE3_COMPARISON_OSC_COSGAIN
For example, here's the command that I used to change the TST line to 0.085:
val=0.085 && caput H1:SUS-ETMX_L3_CAL_LINE_CLKGAIN $val && caput H1:SUS-ETMX_L3_CAL_LINE_SINGAIN $val && caput H1:SUS-ETMX_L3_CAL_LINE_COSGAIN $val && caput H1:CAL-CS_TDEP_SUS_LINE3_COMPARISON_OSC_CLKGAIN $val && caput H1:CAL-CS_TDEP_SUS_LINE3_COMPARISON_OSC_SINGAIN $val && caput H1:CAL-CS_TDEP_SUS_LINE3_COMPARISON_OSC_COSGAIN $val
Although I saw Louis ACCEPT the SDF Diffs for SUSETMx, I think with the safe.snap stuff afterward, maybe it didn't take?
Anyway, We had SUSETMX diffs again. So I ACCEPTED the 9-diffs (attached). Hopefully good for next lock.
Something happened that reverted my changes for CAL-ETMX back to the old settings. Maybe something happened with SDF? Corey brought us out of OBSERVE for me to restore the new amplitude heights. Here's the command I used:caput H1:SUS-ETMX_L3_CAL_LINE_CLKGAIN 0.085 && caput H1:SUS-ETMX_L3_CAL_LINE_SINGAIN 0.085 && caput H1:SUS-ETMX_L3_CAL_LINE_COSGAIN 0.085 && caput H1:SUS-ETMX_L2_CAL_LINE_CLKGAIN 9 && caput H1:SUS-ETMX_L2_CAL_LINE_COSGAIN 9 && caput H1:SUS-ETMX_L2_CAL_LINE_SINGAIN 9 && caput H1:SUS-ETMX_L1_CAL_LINE_CLKGAIN 6.6 && caput H1:SUS-ETMX_L1_CAL_LINE_COSGAIN 6.6 && caput H1:SUS-ETMX_L1_CAL_LINE_SINGAIN 6.6output:Old : H1:SUS-ETMX_L3_CAL_LINE_CLKGAIN 0.3 New : H1:SUS-ETMX_L3_CAL_LINE_CLKGAIN 0.085 Old : H1:SUS-ETMX_L3_CAL_LINE_SINGAIN 0.3 New : H1:SUS-ETMX_L3_CAL_LINE_SINGAIN 0.085 Old : H1:SUS-ETMX_L3_CAL_LINE_COSGAIN 0.3 New : H1:SUS-ETMX_L3_CAL_LINE_COSGAIN 0.085 Old : H1:SUS-ETMX_L2_CAL_LINE_CLKGAIN 50 New : H1:SUS-ETMX_L2_CAL_LINE_CLKGAIN 9 Old : H1:SUS-ETMX_L2_CAL_LINE_COSGAIN 50 New : H1:SUS-ETMX_L2_CAL_LINE_COSGAIN 9 Old : H1:SUS-ETMX_L2_CAL_LINE_SINGAIN 50 New : H1:SUS-ETMX_L2_CAL_LINE_SINGAIN 9 Old : H1:SUS-ETMX_L1_CAL_LINE_CLKGAIN 35 New : H1:SUS-ETMX_L1_CAL_LINE_CLKGAIN 6.6 Old : H1:SUS-ETMX_L1_CAL_LINE_COSGAIN 35 New : H1:SUS-ETMX_L1_CAL_LINE_COSGAIN 6.6 Old : H1:SUS-ETMX_L1_CAL_LINE_SINGAIN 35 New : H1:SUS-ETMX_L1_CAL_LINE_SINGAIN 6.6
OK, for this current lock, we had SDF Diffs for SUSETMX (I'm assuming for the previous lock, we did not save the appropriate .snap file (SAFE.snap or OBSERVE.snap?)).
So for this current lock, the OBSERVE segment from 0222-0238 was with the WRONG (9) SUSETMX CAL_LINE channels & I ACCEPTED these diffs to get to OBSERVE.
Then Louis noticed my alog, and saw that we had the OLD values for these (9) channels! So we immediately ended that OBSERVE segment and worked on correcting SUSETMX. I'm hoping I am doing this right, but this is what we did for the SUSETMX SDF:
- Louis reverted the values to the (9) channels to what they were supposed to be (see his comment above).
- This brings up SUSETMX SDFs.
- I ACCEPTED these.
- Then I clicked SUSETMX's "SDF RESTORE SCREEN" (orange button)
- On new window, I clicked "! SELECT REQUEST FILE" (black button)
- This opens a gray file window for h1susetmx burt files (.snap files)
- I selected the safe.snap file & clicked OPEN (closing this window)
- I clicked the "LOAD TABLE" button on the "SDF RESTORE SCREEN" window
- I believe at this point, the SUSETMX sdf screen went from green to light blue, becaue now the TABLE FILE NAME is the safe.snap (table is filled with all the safe.snap files)---so we can't go to Observe.
- Went back to the "SDF RESTORE SCREEN", and clicked "! SELECT REQUEST FILE"
- Selected the OBSERVE.snap file & clicked OPEN
- Clicked LOAD
- And then SUSETMX returned to green and the OBSERVE file was now the "TABLE FILE NAME"
I think that covers everything. I was able to go back to OBSERVE. Fingers crossed that for the next lock, SUSETMX has all the CORRECT values for these pesky (9) CAL_LINE channels! Sorry for not being more diligent on ACCEPTING diffs! I should have caught the values---I'm glad Louis was online and noticed this from my earlier alog comment!
The reason why Louis' CAL-ETMX gain changes "didn't take" is because they are set by the ISC_LOCK guardian in several places during lock acquisition, including when we reach Nominal Low Noise. This was discovered somewhat recently by Jeff; see his alog 71710 for the full breakdown of the tug-of-war that guardian plays with these gains. Corey and Louis had correctly accepted the new gain values in both the SAFE and OBSERVE SDF tables, but this caused SDF diffs to appear once we reached NLN, preventing us from going into observation.
Since I don't have an answer for how these gains should be set during full lock acquisition, I've just updated the gain values to Louis' new ones for when we reach NOMINAL_LOW_NOISE, not in PREP_FOR_LOCKING or anywhere else. This is set in the gen_NOMINAL_LOW_NOISE() function within ISC_GEN_STATES.py
EDIT: ISC_LOCK has been loaded.