Last week we made a change to the ETMY ring heater to try and bring the arm modes closer together by monitoring the locations of the 2nd order peaks in the 64kHz DCPD channel. The aim was to improve common mode rejection and also reduce the amount of TEM00 constrast defect. Before this, the contrast defect was around 4.8mW (67305) and gave a homodyne angle of 27 degrees. With new RH changes it looks like we have something more like 1mW or ~11 deg homodyne angle. Current RH settings:
| ITMX | 0.44 W |
| ITMY | 0 W |
| ETMX | 0.4 W |
| ETMY | 0.4 W |
The ETMY ring heater was changed early last week from 1.4W -> 0.4W. This moved the 2nd order HOM peaks up from around 10.3kHz to 10.6kHz (Left peak is the Y arm modes). Somewhere in between this the arm buildups looked a better (PRG 52.5 and arm power trending to 390kW) but after reaching a steady state they are dropped to somewhat similar values we had before. Hard to tell where the good point is as we didn't lock again for awhile. Overall we managed to bring the two arm 2nd order peaks to 10.5 and 10.75 kHz. It was at this stage that various ASC locking issues began which took up a large part of last week. We were not able to get any locks long enough to reach a proper steady state and take some measurements until late last week.
During the locking issues, I increased the ETMX RH from 0->0.4W to reduce the mode separation frequency and bring the arms together and in a vain hope it might help locking - it didn't really. Once the ASC issues were solved enough we were having long locks I managed to remeasure the constrast defect using the labutils/darm_offset_scan scripts. These scripts make sure to enable the DARM notches to correctly measure the optical gain change at 255 and 410 Hz. with 0.4W on both ETMs we see about 0.9-1 mW of TEM00 contrast defect (depending on when I measured it during the thermalisation of the lock). This should drop our homodyne angle down to around 11-12 degrees, instead of 27 degrees with 4.8mW.
As the locks started looking reasonable lengths I tried moving ETMX RH further to 0.8W. However, we eventually lost lock and then ran into new 14.7 kHZ parametric instability (68165) problems. It took awhile on Sunday after realising this and gradually dropped the ETMX RH power back down to 0.6W and then finally to 0.4W and we started locking again.
We can tweak the RHs some more and probably reduce this contrast defect further, but it might be worthwhile checking if we see any squeezing gains by going from 27->11 deg homodyne.
Taking the DARM offset scans from Feb 8th (67305) and comparing those from this ring heater change we can see how the contrast defect has dropped and the optical gain has improved at both 255 and 410 Hz.
From this we can plot how much the optical gain has improved as function of chosen DCPD before and after the ring heater change. Running at 20mA we see about a 17% improvement. The improvement from the ring heater change becomes less at higher DC offsets as we're less dominated by the large constrast defect.
I also re-ran Craig's broadband DARM measurement script. In Craig's recent (67963) broadband DARM measurements we can at 100Hz it is about ~6e9 at 20mA, when measured the other day I see ~7e9 = 1.16 improvement. That measurement wasn't done when thermalised so there were some transient effects happening but overall in agreement to the offset scan.
You can also see we have a big optical spring now which needs fixing.
I compared Dan's DARM measurement during the RH move from March 26 2023 a nominal one from February 24, 2023. This is in the same vein as other DARM comparisons As Dan said, there is an across-the-board improvement in optical gain of around 13%. At low frequencies there is a clear DARM spring with the -175 SRCL offset. The phase of the DARM response seems to indicate the DARM spring is below 6 Hz at least, since there's no RHP poles causing the phase to rotate up from -180 to 0. The effect of the ASC on DARM is still dominant in this region. The output of thermal_state_capture at these two times: $ thermal_state_capture 1361329950 and $ thermal_state_capture 1363918073 --------------------------------------------------------------------------------------------------------------- Thermal state of LIGO Hanford at 2023 Feb 25 03:12:12 UTC (1361329950) vs 2023 Mar 27 02:07:35 UTC (1363918073) --------------------------------------------------------------------------------------------------------------- ISC LOCK state number ISC LOCK guardian state number H1:GRD-ISC_LOCK_STATE_N : 700.000 cts vs 700.000 cts Input power PSL requested input power H1:IMC-PWR_IN_OUT16 : 59.406 W vs 59.264 W Input power incident on PRM H1:IMC-IM4_TRANS_NSUM_OUT16 : 56.614 W vs 56.578 W PRG Power recycling gain H1:LSC-PR_GAIN_OUT16 : 49.583 W/W vs 51.057 W/W Reflected power LSC REFL A DC power H1:LSC-REFL_A_LF_OUT16 : 6.454 mW vs 7.632 mW LSC REFL B DC power H1:LSC-REFL_B_LF_OUT16 : 6.351 mW vs 7.501 mW POP power LSC POP A DC power H1:LSC-POP_A_LF_OUT16 : 29673.473 uW vs 30468.768 uW POP 2f power POPAIR B I 18 MHz H1:LSC-POPAIR_B_RF18_I_OUT16 : 1493.631 cts vs 1497.223 cts POPAIR B Q 18 MHz H1:LSC-POPAIR_B_RF18_Q_OUT16 : -2657.690 cts vs -2142.292 cts POPAIR B I 90 MHz H1:LSC-POPAIR_B_RF90_I_OUT16 : 409.637 cts vs 363.837 cts POPAIR B Q 90 MHz H1:LSC-POPAIR_B_RF90_Q_OUT16 : -658.049 cts vs -556.946 cts DARM offset DARM offset setting (RF DARM) H1:LSC-DARM1_OFFSET : 0.000 cts vs 0.000 cts DARM offset setting (DC readout) H1:OMC-READOUT_X0_OFFSET : 7.700 cts vs 7.739 cts DCPD SUM milliamps H1:OMC-DCPD_SUM_OUT16 : 19.792 mA vs 20.005 mA RF mod depth sliders 9 MHz modulation depth slider H1:LSC-MOD_RF9_AM_RFSET : 20.400 dBm vs 20.400 dBm 45 MHz modulation depth slider H1:LSC-MOD_RF45_AM_RFSET : 21.000 dBm vs 21.000 dBm Ring Heaters ETMX ring heater upper power H1:TCS-ETMX_RH_UPPERPOWER : 0.000 W vs 0.388 W ETMX ring heater lower power H1:TCS-ETMX_RH_LOWERPOWER : 0.000 W vs 0.391 W ITMX ring heater upper power H1:TCS-ITMX_RH_UPPERPOWER : 0.429 W vs 0.429 W ITMX ring heater lower power H1:TCS-ITMX_RH_LOWERPOWER : 0.426 W vs 0.426 W ETMY ring heater upper power H1:TCS-ETMY_RH_UPPERPOWER : 1.368 W vs 0.389 W ETMY ring heater lower power H1:TCS-ETMY_RH_LOWERPOWER : 1.367 W vs 0.390 W ITMY ring heater upper power H1:TCS-ITMY_RH_UPPERPOWER : 0.000 W vs 0.000 W ITMY ring heater lower power H1:TCS-ITMY_RH_LOWERPOWER : 0.000 W vs 0.000 W CO2s CO2X power monitor H1:TCS-ITMX_CO2_LSRPWR_MTR_OUT16 : 3.977 W vs 3.988 W CO2Y power monitor H1:TCS-ITMY_CO2_LSRPWR_MTR_OUT16 : 1.707 W vs 1.705 W A2L Gains ETMX pitch-to-length spot H1:SUS-ETMX_L2_DRIVEALIGN_P2L_SPOT_GAIN : 4.000 cts vs 4.000 cts ETMX yaw-to-length spot H1:SUS-ETMX_L2_DRIVEALIGN_Y2L_SPOT_GAIN : 4.400 cts vs 4.400 cts ETMY pitch-to-length spot H1:SUS-ETMY_L2_DRIVEALIGN_P2L_SPOT_GAIN : 4.600 cts vs 4.600 cts ETMY yaw-to-length spot H1:SUS-ETMY_L2_DRIVEALIGN_Y2L_SPOT_GAIN : 3.200 cts vs 3.200 cts ITMX pitch-to-length spot H1:SUS-ITMX_L2_DRIVEALIGN_P2L_SPOT_GAIN : -0.800 cts vs -0.800 cts ITMX yaw-to-length spot H1:SUS-ITMX_L2_DRIVEALIGN_Y2L_SPOT_GAIN : 2.100 cts vs 2.100 cts ITMY pitch-to-length spot H1:SUS-ITMY_L2_DRIVEALIGN_P2L_SPOT_GAIN : -0.050 cts vs -0.050 cts ITMY yaw-to-length spot H1:SUS-ITMY_L2_DRIVEALIGN_Y2L_SPOT_GAIN : -1.700 cts vs -1.700 cts TSAMs TSAMs (OM2) thermistor 1 temperature H1:AWC-OM2_TSAMS_THERMISTOR_1_TEMPERATURE : 22.283 Celcius vs 25.268 Celcius TSAMs (OM2) thermistor 2 temperature H1:AWC-OM2_TSAMS_THERMISTOR_2_TEMPERATURE : 22.636 Celcius vs 38.427 Celcius
May I ask a question about the reduction of homodyne angle (contrast defect)?
I understand that the contrast defect light is reduced because the arm modes are better matched to each other after optimizing ring heaters. However, I would also expect less HOM light being outputed to the AS port and thus less light rejected by OMC on the OMC REFL PD. It looks like the OMC REFL didn't see much difference after changing the ETMY ring heater. The DARM offset and RF 9/45 modulation index were unchanged during this time.
I guess my ultimate question is where the large contrast defect light (~ 4.8 mW) was from. If RH helps reduce it, then it's likely that it's mostly from the arm mode mismatch. How much would the mismatch be to produce 5 mW contrast defect? I tried 1% (which is pretty large I think), but it's not enough. Thanks!
