[Haoyu Wang, Yuta Michimura] (Measurement done on Aug 14)
We measured the noise spectra of TMSX and TMSY polarization monitors while single Xarm and Yarm are locked.
Measured polarization rotation angle of TRX and TRY are currently ~1e-5 rad/rtHz above 10 Hz, limited by ADC noise.
Method:
- Locked single Xarm (Yarm) with PRM misaligned, at IMC output power of ~1.1 W.
- Measured the spectrum of following channels.'K1:TMS-%s_IR_PDA1_OUT_DQ' % ARM
'K1:TMS-%s_IRSPOL_PDA1_OUT_DQ' % ARM
'K1:TMS-%s_IRPPOL_PDA1_OUT_DQ' % ARM
- Converted into polarization rotation angle using the calibration done for TRX (klog #30113) and TRY (klog #30827) using the following. Note that HWP angle was 20 deg for Xarm (we set it on Aug 13) and 80 deg for Yarm (klog #30827).if ARM=='X':
startTime=1407634590 # 1.1 W at IMC output, whitening off, Xarm
PDs2Rad=1/(4*(297.0-19.7)*np.sin(4*np.deg2rad(20-54.7)))
elif ARM=='Y':
startTime=1407636370 # 1.1 W at IMC output, whitening off, Yarm
PDp2Rad=1/(4*(358.0-24.1)*np.sin(4*np.deg2rad(80-72.5)))
PDs2Rad=1/(4*(352.0-24.2)*np.sin(4*np.deg2rad(80-117.4)))
Results:
- Attachment #1 (for X) and #2 (for Y) are the raw spectra for the channel. K1:TMS-X_IR_PDA1_OUT_DQ seems less noisy, as K1:TMS-X_IR_PDA1_PDA1_GAIN=0.25504554. Note that K1:TMS-Y_IR_PDA1_PDA1_GAIN=1.19592. All the channels are basically limited by ADC noise above ~10 Hz.
- Attachment #3 (for X) and #4 (for Y) are the polarization rotation angle estimated from the acquired data.
Next:
- Change PD gains and whitening gains after locking PRFPMI so that the spectrum will not be limited by ADC noise. Calibrate the PDs again.
Michimura, Takano
We tried to measure the spectra of TMSX and TMSY poilarization monitor with PRFPMI. It seems that the monitor PDs for s-pol are saturated, but for p-pol seems to work.
PD saturation check
Figure1 shows the raw spectra and coherence of TMSX and TMSY polarization monitor. It is apparent that the spectra of X arm s-pol and Y arm s-pol are not sensitive to the power fluctuation. I checked the time series data as shown in Figure2, and found that the output of these two PDs are saturated.
Referring to the power measurement, the power injected to the polarization monitors and the corresponding output DC voltage are as listed below. I also estimated the power and output with PRFPMI so that the equivalent input power is 140 times larger than the previous measurement, single arm lock. In this calculation the response of the monitor PDs are taken from this post, 0.39 [A/W] and 4.75e4 [V/A] for 30dB gain setting.
| Single arm | PRFPMI | ||||
| Power [W] | Output [V] | Power [W] | Output [V] | ||
| TMSX | s-pol | 16.4e-6 | 0.304 | 2.3e-3 | 43 |
| p-pol | 2.0e-6 | 0.037 | 0.28e-3 | 5.2 | |
| TMSY | s-pol | 18.0e-6 | 0.333 | 2.5e-3 | 47 |
| p-pol | 2.4e-6 | 0.044 | 0.33e-6 | 6.2 | |
From the calculation it is confirmed that with PRFPMI the s-pol monitor PDs should be saturated. To avoid the saturation, we need to reduce the gain setting of s-pol monitor PDs from 30dB to 10dB (4.75e3 [V/A]) or lower. When the input power is increased to 10 W, we also have to reduce the gain of p-pol monitors.
Polarization rotation angle
Using p-pol monitor signal, we estimated the polarization rotation angle.
startTime=1408765218 # 1.1 W at IMC output, whitening off- Calibration: almost same as the original post , but 130 time smaller (because of increased transmission power)
Figure3 (Figure4) and Figure5 (Figure6) shows the spectra of polarization monitors of TMSX (TMSY) in cnt/rtHz and the polarization rotation angle of TMSX (TMSY) in rad/rtHz. Now the sensitivity is roughly 100 times better than the previous measurement.
[Marc Eisenmann, Satoru Takano, Yuta Michimura]
We lowered the PD gain for TMS Spol PD by a factor of 10, and turned on two whitening stages for Spol and Ppol PDs, for both Xend and Yend.
TMS polarization monitor PDs are not limited by ADC noise any more, and are sensitve to 1e-8 rad/rtHz level.
What we did:
- Lowered PD gains from 30 dB to 10 dB for IRSPOL PDs for both TMSX and TMSY, as they were saturating after PRFPMI lock (klog#30892)
- Turned on analog whitening filters and digital anti-whitening filters one by one to see how many stages are necessary to reduce ADC noise. We confirmed that 2 stages are enough (See Attachment #1 for X and Attachment #2 for Y). Spectra for Y somehow have more structures than those for X.
- Calibrated measured spectra into polarization rotation angle using the factor described in klog #30885, but 130 smaller to take into account of power difference between single arm lock and PRFPMI lock, and another 10 times smaller for IRSPOL_PDA1 to take into account of the PD gain change. See Attachment #3 and Attachment #4 for the original data in counts for X and Y, respectively, and Attachment #5 and Attachment #6 for the calibrated data for X and Y, respectively.
- Attachment #7 and Attachment #8 are the photos of the whitening filter switch after the work for X and Y, respectively.
Summary of PD whitening situations (same for TRX and TRY):
IR_PDA1: 0 dB, no whitening gain, no whitening filters
IRSPOL_PDA1: 10 dB, no whitening gain, 2-stage whitening filter
IRPPOL_PDA1: 30 dB, no whitening gain, 2-stage whitening filter
Next:
- Re-calibrate after the power up
- Tune whitening for IR_PDA1
