I made a first integrated noise budget for PRFPMI with NoiseBudgetter. Plot below is from Mar 9, 2020 14:00 UTC when we reached 0.6 Mpc.
See, also klog #11755 for NB for FPMI.
Shot noise estimated based on measurements done with lower power suggests we are mostly limited by shot noise above 200 Hz.
NoiseBudgetter configuration file lives in
/kagra/Dropbox/Subsystems/MIF/NoiseBudget/PRFPMI/NPconf_20200311_1119.csv
Details
* Suspension thermal noise
- Suspension thermal noise of cryogenic payload at 300 K, based on Somiya-san's calculation (300Knew.nb)
* Mirror thermal noise
- Substrate + thermoelastic + coating thermal noise of sapphire mirror at 300 K, based on Somiya-san's calculation (300Knew.nb)
* Shot noise
- Based on Enomoto-kun's calculation in klog #12772. It was roughly 1.3e-19*sqrt(1+(freq/22 Hz)**2) /rtHz. I simply scaled this with sqrt(1.4/Pin)*sqrt(3.4mW/8mW), where Pin is the IMC output power (2.9 W for this plot). [NEED TO REDO THIS WITH CURRENT OPTICAL GAIN]
- Cavity pole assumed to be 22 Hz, which is the number I heard. [NEED CONFIRMATION]
* OMC DC PD dark noise
- Dark noise of OC DC PD measured at K1:CAL-CS_PROC_DARM_DISPLACEMENT_DQ (there fore you have to calibrate it with 5-stage dewhitening to make it actual DARM displacement spectrum as usual). I'm not sure the details, but DARMsens.xml says it is measured on Feb.11 with 1stage wh.
* Type A noise
- Noise from klog #13474. Sum of all four suspensions.
* Intensity noise
- Based on measurement done by me in klog #13028. K1:PSL-ISS_FIRST_SERVO_PDA_RIN_OUT_DQ is projected to DARM displacement with a gain of 1.1e-12 m and zeros at 100 Hz, 2500 Hz.
* Frequency noise
- K1:LSC-CARM_RESIDUAL_OUT_DQ is projected to DARM with 3e-14. I'm not sure where this 3e-14 is from. DARMsens.xml says "based on 2020 Feb 14" [NEED CONFIRMATION]
- Related: klog #13442
Discussion on shot noise
- Shot noise calculation above is a very rough estimate. We should redo the shot noise measurement.
- Also, theoretical quantum noise with 30 W at BS, 70% loss from SRM is a factor of 2.5 smaller (see attached). This might suggest some very big loss. Please measure the optical loss from BS to OMC DC PD (as was done in klog #11762, with AS RF).
Next:
- Include MICH and PRCL projections
- Let me know if you have your favorite noise projections. Let me know the name of the witness channel, and its coupling transfer function to DARM.
I realized that BS to OMC DC PD throghput is measured today in klog #13476.
3 W at IMC output means 3 W * 10.35 % * 50 % * 50 % = 77.6 mW at BS output to SRs.
So, BS to OMC DC PD throughput is 3.3 mW / 77.6 mW = 4.3 %.
The updated theoretical quantum noise using this value is attached (see red dotted line).
Green curve is the measured shot noise in klog #12772, scaled with current input power and power at OMC DC PD.
They are consistent within 10 %.
I have updated NB for PRFPMI on Mar 9, 2020.
The NoiseBudgetter configuration file lives in
/kagra/Dropbox/Subsystems/MIF/NoiseBudget/PRFPMI/NPconf_20200314_1836.csv
Details:
* Shot noise
- Updated to use Aso calculation in klog #13475. Cavity pole was set to 22.2 Hz (previously 22 Hz). I got this number from DARM reconstruction filter (from FM2 filter in K1:CAL-CS_PROC_DARM_FILT_INVC).
- Modeled with 7.4e-20*(1+(freq/cavity_pole)**2)**0.5.
- Shot noise based on Enomoto-kun's calculation scaled with current power (see klog #13481), and theoretical calculation based on (see klog #13485), and Aso-san's measurement agree within 30% (see attached).
* MICH and PRCL coupling
- I used /users/yokozawa/Data/PRFPMI/MICH_PRCL_TO_DARM/200312/TF_MICH_DARM_200312.txt and /users/yokozawa/Data/PRFPMI/MICH_PRCL_TO_DARM/200312/F_PRCL_DARM_200312.txt are used for coupling function and K1:LSC-MICH_IN1_DQ and K1:LSC-PRCL_IN1_DQ as witness channels.
- see klog #13518
* Type A controls noise
- I didn't updated this but someone please!
- see klog #13508
I also added comments to each noise (Thank you Shoda-san for the bug fix!)
Michimura-san
Does it include the DAC noise of the ETMs?
As part of the double check for the shot-noise level, I have also made yet another estimate for the shot-noise-limited sensitivity.
The conclusion: my estimate also agrees with those derived by Yoichi [13475] and Yuta [13560].
The shot noise level reaches appoximately 3x10-18 m/Hz1/2 at 1 kHz, fully consistent with Yoichi's and Yuta's estimations.
Some more details can be found at JGW-T2011554-v1.
Some side products:
- The size of the DARM offset is estimated to be Lx-Ly = 17.1 pm when Pin=3.1 W and DCPD A = 8 mA.
- The DARM optical gain is estimated to be dP/d(Lx-Ly) = 1.09 mW/pm at DC when Pin =3.1 W and DCPD A = 8 mA.
Today I estimated the DAC noise at TM stage as following:
DAC_noise_TM=TF_TM_L2Darm*c
where c is the measured DAC noise value and it is equal to 4.5 e-3 counts/rHz. (see this ref. )
The attached plots show the results of this calculation:
-Pic1 shows the calibrated TFs from TM_L to DARM for all TypeA: The TFs show same behavior and can be fitted as 1/f^2.
-Pic2 shows the DAC noise projection measured at IX (yellow line), EY (red line) and the total DAC noise contribution:
According with this plot seems that around 100 Hz we are limited by the ETMs DAC noise.
I have updated NB for PRFPMI on Mar 9, 2020 to include recent suspension related noises.
The NoiseBudgetter configuration file lives in
/kagra/Dropbox/Subsystems/MIF/NoiseBudget/PRFPMI/NPconf_20200316_1040.csv
Details:
* Type A controls noise
- Local damping feedback signal calibated to DARM (basically sensor noise).
- Sum of controls noises in klog #13590
* Type A DAC noise
- DAC noise from actuators at TM stages of Type-A suspensions.
- Modeled with 8e-14/freq**2
- Sum of DAC noises in klog #13589
* BS controls noise
- Local damping feedback signal calibated to DARM (basically sensor noise).
- Sum of controls noises in klog #13587
Thank you Lucia and Fabian for presenting noise projections in the NoiseBudgetter format!
I realized that in the calculation of the total DAC noise, I overstimated it by a factor 2 (klog 13589).
In the previus calculation the DAC noise contribution for each TM was estimated as following:
1) DAC_noise_TM=TF_TM_L2Darm*c
-In this espression I didn't take into account that the DAC noise pushes 4 coils incoerently. If we consider this, the corrected formula should be:
2) DAC_noise_TM=1/2* TF_TM_L2Darm*c
- I didn't take into account the effect of the DW filters engaged at ITMs.
By taking into account the DW filter attenuation (about a factor 10 at 10 Hz), eq2 can be written as:
3) DAC_noise_TM_dw=1/(2*10)* TF_TM_L2Darm*c
Considering this, today I estimated the DAC noise at TM stage by using eq 2 for ETMs and eq 3 for ITMs.
The attached plots show the results of this calculation:
-Pic1 shows the DAC noise projection measured at IX (DW filter), EY (w/o DW filter) and the total DAC noise contribution (sqrt (IX^2+IY^2+EY^2+EX^2))
According with this plot seems that around 100 Hz we are limited by the ETMs DAC noise.
-Pic2 shows the DAC noise projection when the DW filters are engaged to IX and EY, the total DAC noise contribution when the DW filter are used on all TMs and the total DAC when the DW filter are turned on only at ITMs.
According with this plot the total DAC noise will be reduced by a factor 10, by turning on the DW filters at ETMs.
*Note
I am assuming that the IY and EX DAC noise contributions should be equal to IX and EY ones respectively because the IY TF from TM_L to DARM is equal to IX ones and the EX TF from TM_L to DARM is equal to EY ones. (see Pic1 in klog 13589)
As I pointed out in today's comissoning meeting, the unknown peaks around 70-100 Hz might be coming from the vibration of bellows connected to BS. (http://klog.icrr.u-tokyo.ac.jp/osl/?r=13597)
I want to put on some accelerometers on them and check it, if i get a chance.
By the way, now I'm trying to include the PEM information to the NoiseBudgetter. Please wait a bit.
I want to put on some accelerometers on them
Me, too!!
I putted the portable accelerometer and compared the spectrum with and without hammering.
Without hammering, the spectrum is quite white(under the sensor noise)
With hammering, we can see the similar spectrum detected in DARM.
I suspect similar scattering happend betwen PR2 and BS.
silent :
Hammering :
I introduced the PEM projection datapoints to NB.
The .txt files are
- /users/PEM/NoiseBudgetter/Sound_PRMbooth_2020-03-07.txt
- /users/PEM/NoiseBudgetter/Sound_BSbooth_2020-03-11.txt
