We are analyzing the data taken with different modulation indices described in alog 58946 to get the different RF field components at each detector port. The analysis uses the known modulation indices (alogs 41889 and 41435) and their change to deduce the amount of 9 MHz RF sidebands, 45 MHz RF sidebands and carrier. A previous determination before the ITM was swapped can be found in alog 48595. See the attached nb file for details.
The table below lists the fractional contribution of each RF component in units of percent power. The columns with the tag "dec" were measured by decreasing the modulation depth, whereas "inc" indicates increased modulation depth. The precision is 0.1% at best, but the AS port measurements are far less consistent.
| Port | DC | 45 MHz dec | 45 MHz inc | 9 MHz dec | 9 MHz inc | Carrier dec | Carrier inc |
|---|---|---|---|---|---|---|---|
| Input power (W) | 45.8 | 1.66 | 1.67 | 0.87 | 0.87 | 97.5 | 97.5 |
| REFL_A (mW) | 18.1 | 4.3 | 4.3 | 5.7 | 5.7 | 90.0 | 90.0 |
| REFL_B (mW) | 16.7 | 4.3 | 4.2 | 5.6 | 5.7 | 90.1 | 90.1 |
| POP_LF (cnt) | 50725 | 0.17 | 0.22 | 1.7 | 1.7 | 98.1 | 98.1 |
| POP_A_NSUM (cnt) | 6045 | 0.20 | 0.34 | 1.6 | 1.6 | 98.2 | 98.0 |
| POP_B_NSUM (cnt) | 5493 | 0.24 | 0.29 | 1.7 | 1.7 | 98.0 | 98.1 |
| AS_C_NSUM (mW) | 320 | 85.7 | 84.0 | 1.7 | 1.6 | 12.6 | 14.5 |
| OMC_A_NSUM (mW) | 305 | 87.7 | 85.9 | 1.6 | 1.5 | 10.7 | 12.7 |
| OMC_B_NSUM (mW) | 314 | 90.5 | 88.6 | 1.5 | 1.4 | 8.0 | 10.0 |
| TR_X_A (kcnt) | 270 | <0.2 | <0.2 | <0.2 | <0.2 | 100 | 100 |
| TR_X_B (kcnt) | 352 | <0.2 | <0.2 | <0.2 | <0.2 | 100 | 100 |
| TR_Y_A (kcnt) | 364 | <0.2 | <0.2 | <0.2 | <0.2 | 100 | 100 |
| TR_Y_B (kcnt) | 385 | <0.2 | <0.2 | <0.2 | <0.2 | 100 | 100 |
The measurements on the input power are consistent with the estimated modulation indices. It is unclear why the 45 MHz contributions measured at the dark port vary by up to 5%.
Performing the same analysis for the April 2019 measurements we get the table below. Back then, only a modulation depth increase was performed.
| Port | DC | 45 MHz | 9 MHz | Carrier |
|---|---|---|---|---|
| Input power (W) | 35.4 | 1.67 | 0.88 | 97.5 |
| REFL_A (mW) | 5.3 | 14.0 | 3.6 | 82.4 |
| REFL_B (mW) | 4.9 | 13.9 | 3.6 | 82.5 |
| POP_LF (cnt) | 32137 | 0.36 | 1.2 | 98.5 |
| POP_A_NSUM (cnt) | 3871 | 0.39 | 1.2 | 98.5 |
| POP_B_NSUM (cnt) | 3555 | 0.40 | 1.2 | 98.4 |
| AS_C_NSUM (mW) | 245 | 70.0 | 9.0 | 21.4 |
| OMC_A_NSUM (mW) | 226 | 75.3 | 8.0 | 16.7 |
| OMC_B_NSUM (mW) | 221 | 80.1 | 6.3 | 13.6 |
| TR_X_A (kcnt) | 173 | <0.1 | <0.1 | 100 |
| TR_X_B (kcnt) | 226 | <0.1 | <0.1 | 100 |
| TR_Y_A (kcnt) | 247 | <0.1 | <0.1 | 100 |
| TR_Y_B (kcnt) | 262 | <0.1 | <0.1 | 100 |
Finally, we are going to compare the new and old measurements by first scaling for the input power and then taking the power ratios of new over old.
| Port | DC | 45 MHz | 9 MHz | Carrier |
|---|---|---|---|---|
| Input power | 1 | 1 | 0.98 | 1 |
| REFL_A | 2.64 | 0.81 | 4.2 | 2.88 |
| REFL_B | 2.65 | 0.81 | 4.2 | 2.89 |
| POP_LF | 1.22 | 0.74 | 1.77 | 1.22 |
| POP_A_NSUM | 1.21 | 1.07 | 1.69 | 1.20 |
| POP_B_NSUM | 1.20 | 0.85 | 1.71 | 1.19 |
| AS_C_NSUM | 1.00 | 1.22 | 0.29 | 0.68 |
| OMC_A_NSUM | 1.04 | 1.19 | 0.24 | 0.79 |
| OMC_B_NSUM | 1.10 | 1.22 | 0.24 | 0.81 |
| TR_X_A/B | 1.21 | — | — | 1.21 |
| TR_Y_A/B | 1.14 | — | — | 1.14 |
Conclusions from these measurements:
- The carrier power recycling gain has increased by a factor of 1.2.
- The 9MHz power recycling gain has increased by a factor of ~1.7.
- The total power in reflection has increased by 2.65 (shot noise x1.6), whereas the 9 MHz sidebands have increase by ~4.18 (signal x2.05). With the carrier gain in the PRC being also 1.2 times higher (signal 1.1), this is yielding an improvement of the REFL SNR of about 1.4 at equivalent input power, or ~1.6 with the increased input power, whereas the signal gain is about 2.2 for equivalent input power.
- With an input efficiency from PSL to PRM of 88.8%, a loss in the REFL path of 28.4%, with four 50:50 and one 90:10 beamsplitters, the overall efficiency at the REFL port LSC PDs is ~1/223 (see alog 46386). This yields an estimate that about ~9% of the incident carrier power is showing up at the reflected port, ~25% of the 45MHz sidebands, and ~65% of the 9 MHz sidebands. This greatly reduced the mystery what happened to the 9 MHz power—with the point absorber on the old ITM being the main culprit for the lousy 9MHz build-up in the past. Still, without losses we would expect close to 100% of the 9 MHz power to show up in reflection.
- At the AS port, the total power stayed about the same for equivalent input power with the carrier contribution decreasing by ~0.7-0.8, the 45 MHz contribution increasing by ~1.2, and the 9 MHz contribution reducing by a factor of ~4-5.5. With the increase of the 9 MHz sideband gain of ~1.7 in the power recycling cavity, the 9 MHz contrast defect has improved by roughly 7-10!
- From the estimated 320 mW at the AS port, 269 mW are 45 MHz sidebands, 32-46 mW are carrier and 4-5 mW are 9 MHz sidebands. With 23.3 mW of the carrier in the TEM00 mode, we end up with 9-23mW in carrier higher order modes. This is significantly better than with the old ITM, where the carrier TEM00 and HOM powers were about the same, even at a lower input power. Estimating 2.2 kW of carrier light at the beamsplitter, the carrier contrast defect is now about 13-32 ppm, where we had 20-66 pm previously. However, 87 ppm were reported in alog 56590. The uncertainties are due to the rather inconsistent measurements coming from ASC-AS_C, OMC_A and OMC_B photodetectors.
- The power gain in in the Y-arm is consistently lower than the X-Arm. This is due to the change of the ITMY transmission that changed from 1.50% to 1.42%, which results in a 0.95 reduction in the power build-up.
