craig.cahillane@LIGO.ORG - posted 18:48, Monday 12 September 2022 - last comment - 16:04, Monday 26 September 2022(64959)
Intensity noise to DARM at low frequencies from August 18 2022
I processed some bandlimited intensity noise injections I did on August 18 of this year. The goal was to resolve the low-frequency intensity noise projection to DARM. This is a difficult business because the intensity loop works strongly suppresses low frequency noise, and easily saturates when you inject a high levels to see the noise in DARM. In our case, we've seen that increasing the intensity injection amplitude actually reduces the coherence seen, since nonlinear noise rises faster with higher intensity RMS. So to reduce RMS, I made some fine bandlimited injections, saved in the same place as alog 52864. The results are strange, quite unlike our intensity to DARM results from O3. See plot 1 for the TF in m/RIN, plot 2 for the DARM projection in m/rtHz. First, the total coupling is far lower now than in the past at LF, see alog 52864. At HF, it's about the same. Second, there is some unexpected dips between 20 and 30 Hz. I sort of expected this region to be dominated by differential radiation pressure due to arm power differencesdP_arms. Third, there is a clear dip at 70 Hz, and a corresponding phase flip from around -45 to 45 degs. We saw something similar to this in the past, as we switch from radiation pressure to contrast defect intensity coupling. (There is a sign difference between these plots and alog 52864. I believe the signs in this alog are correct.)------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Full data for these plots is in:/ligo/home/craig.cahillane/Git/IFO/IntensityNoise/data/Injections/20220818Scripts for generating these plots in:/ligo/home/craig.cahillane/Git/IFO/general/scriptsTerminal calls were:To process data (in labutils env): python run_all_intensity_noise_injection_plots.py /ligo/home/craig.cahillane/Git/IFO/IntensityNoise/data/Injections/20220818/Intensity_inj_*.hdf5 1344821693 1344831693 -o 0.5 To stitch TFs (in psl env): python /ligo/gitcommon/psl_measurements/code/multi_tf_plot.py *intensity_to_darm_tf.txt --stitch -u --title "Stitched bandlimited intensity to DARM injections - August 18 2022" -y log -s --labels "100-200 Hz" "2000-7000 Hz" "200-300 Hz" "20-30 Hz" "300-2000 Hz" "30-40 Hz" "40-60 Hz" "60-100 Hz" --legend_loc outside -y1 1e-14 -y2 1e-11 --units "m/RIN" To stitch ASDs (in psl env): python /ligo/gitcommon/psl_measurements/code/multi_spec_plot.py *intensity_to_darm_asd_projection.txt --stitch --title "Stitched bandlimited intensity to DARM injections - August 18 2022" -s -x1 10 -x2 7500 -y1 1e-23 -y2 1e-16 --ylabel "DARM [m/rtHz]" --labels "100-200 Hz" "2000-7000 Hz" "200-300 Hz" "20-30 Hz" "300-2000 Hz" "30-40 Hz" "40-60 Hz" "60-100 Hz"
Non-image files attached to this report
Comments related to this report
I measured the CARM pole using these injections, similar to alog 26768. The transfer functions measured are Transmitted Arm RIN / Input RIN. If we assume that the Input RIN is dominated by carrier RIN, then the comparison TF should give us the CARM pole frequency directly. Results------- CARM pole = 0.576 +- 0.009 HzThe 1.6% uncertainty is dominated by the spread of the CARM pole fit values from each QPD, each shown in the PDF. The RIN-ification process for these channels is somewhat arbitrary: I divided the arm trans channels (H1:ASC-{X,Y}_TR_{A,B}_NSUM_OUT) by a median value. The power on the PD drifts around as the TMS drifts and the interferometer thermalizes, but during the measurement windows it's fairly stable.CodeFirst PDF made by/ligo/home/craig.cahillane/Git/IFO/IntensityNoise/scripts/carm_pole_from_input_rin_to_ifo_rin_tf.pyThe rest made by/ligo/home/craig.cahillane/Git/IFO/general/scripts/general_bandlimited_noise_injections_processor.py
Non-image files attached to this comment
Posting input RIN TF results to many other ports of the interferometer as well, for future musings. 1) REFL A LF RIN 2) REFL B LF RIN 3) AS_C RIN 4) REFL A 9 I 5) REFL B 9 I 6) REFL A 45 I 7) AS A 45 I Some notes: 1) REFL LF RIN is sort of expected to be higher, because the intensity noise is rejected from entering the interferometer, but the main beam is largely accepted, so we increase the RIN by dividing by a smaller carrier in reflection. 2) AS_C RIN is dominated by 45 RIN. 3) Looking at REFL 9 I, we do witness frequency noise as a result of intensity noise. Before, we measured that intensity noise is not witnessed as frequency noise in reflection of the IMC by itself alog 62240, but did not check in transmission of the IMC. So this intensity-to-frequency coupling could originate from IMC transmission or IFO reflection. 4) I don't know what to make of the bonkers input RIN to REFL 45 I coupling. It oddly mirrors the DARM coupling dip, but at a different frequency region.
Non-image files attached to this comment