Jennie W, Sheila
Summary: We altered the offsets on the H1:ASC_OMC_{A,B}_{PIT,YAW} QPDs which are used to align the beam into the OMC. This was aiming to give us a improvement in optical gain. After doing this we aimed to measure the anti-symmetric port light changing as we chnage the darm offset. We are trying to use both these measurements to narrow down where we have optical loss in that could be limiting our observed squeezing. Performed both measurments successfully but the different alignment of the OMC made the squeezing less good so Camilla (alog #83009) needed to do some tuning.
Last time (alog #82938) I did this I used the wrong values as our analysis used the output channels to the loops instead of the input channels which come before the offsets are put in. The new analysis of our measurement of the optical gain as seen by the 410Hz PCAL line, changing with QPD offset, shows that we want the loop inputs to change to:
H1:ASC_OMC_A_PIT_INMON to 0.3 -> so we should change H1:ASC_OMC_A_PIT_OFFSET to -0.3
H1:ASC_OMC_A_YAW_INMON to -0.15 -> so we should change H1:ASC_OMC_A_YAW_OFFSET to 0.15
H1:ASC_OMC_B_PIT_INMON to 0.1 -> so we should change H1:ASC_OMC_B_PIT_OFFSET to -0.1
H1:ASC_OMC_B_YAW_INMON to 0.025 - so we should change H1:ASC_OMC_B_YAW_OFFSET to -0.025
We stepped these up in steps of around 0.01 to 0.02 while monitoring the saturations on OMC and OM3 suspensions and the optical gain, both to make sure we were going in the correct direction and that we were not near to saturation of the suspensions as hapenened last time I tried to do this.
Attached is the code and the ndscope showing the steps on each offset, (top row left plot, top row center right plot, second row left plot, second row center right plot). The top stage osems for OM3 suspension are shown in the third row left plot, the top stage osems for OMC suspension are in the third row center left plot, and the optical gain is shown in the third row right plot.
The optical gain improved from by 0.0113731 from a starting value of 1.00595, so that is an improvement of 1.13 % in optical gain.
Around 19:04:28 UTC I started the DARM offset step to see if the change in optical gain matches that we would see if we measured the throughput of HAM 6. Unfortuntely I forgot to turn off the OMC ASC which we know affects this measurement of the loss. We stood down from changing the OMC and Camilla did some squeezer measurements, then I made the same mistake again the next time I tried to run it (d'oh). Both times I control-C'd the auto_darm_offset.py form the command line which means the starting PCAL line values, and DARM offset had to be reset manually before I ran the script successfully after turning the OMC ASC gain to 0 to turn it off.
The darm offset measurement started at 19:20:31 UTC. The code to run it is /ligo/gitcommon/darm_offset_step/auto_darm_offset_step.py
The results are saved in /ligo/gitcommon/darm_offset_step/data and /ligo/gitcommon/darm_offset_step/figures/plot_darm_optical_gain_vs_dcpd_sum.
From the final plot in the attached pdf, the transmission of the fundamental mode light between ASC_AS_C (anti-symmetric port) DCPD is (1/1.139)*100 = 87.8 %. We can compare this to the previous measurement from last week with the old QPD offsets to see if the optical loss change matches what we would expect from such a change in optical gain.
Since the script didn't save the correct values for pcal ey and ex (due to the script being run partially twice before a siccessful measurement). I reverted the PCAL values back using SDF before we went into observing. See attached screenshots.
Sheila accepted the new ASC-OMC_A and B OFFSET values in OBSERVE and SAFE (only have the pic for OBSERVE).
Comparing OMC losses calculated by OMC throughput and optical gain measurements.
If we take the improvement in optical gain noted above and calculate the improvment in the optical gain ^2, ie.
(g_f^2 - g_i^2)/ g_i^2 = 0.023 = 2.3 %
And compare it to the gain in OMC throughput from this entry to the measurement after changing the OMC ASC offsets above
(T_OMC_f - T_OMC_i)/ T_OMC_i = 0.020 = 2%
Both methods show a similar improvement in the coupling to the OMC, or alternatively decrease in the HAM 6 losses. Since we improved the alignment of the OMC, it makes sense that the losses decrease and them agreeing validates our method of using darm offset steps to calculate OMC throughput and thus the loss in HAM 6.
The optical gain must be squared as it changes with the square root of the power at the output (due to the DARM loop).
For this comparison I was not able to use the measurement of optical gain from the same day as the initial measurement of OMC throughput, (alog #82938) as the calibration was exported to the front-end between these two dates which would have changed the reference value for kappa C.
The code I used for calcultions is attached.
As I did for the previous DARM offset measurement on the 20th Feb, in alog #83586, I checked that the DARM offset does not show a clear trend in the OMC REFL power. This would be another way of quantifying the mode-matching of the DARM mode to the OMC, but since the mode-matching is good, no trend can be seen in this channel (top plot) as we change the DARM offset.
