NAOJ GW Elog Logbook 3.2
7/27 [Chien-Ming, Hsun-Chung]
Summary:
We measured about 5 dB squeezing and 11 dB anti-squeezing at 25.1mW pump power.
Current parameter settings:
DDS1 | ch1 | SHG & IRMC demod. | phase 140 deg. |
DDS2 | ch2 | GRMC demod. | phase 200 deg. |
DDS3 | ch3 | P-Pol PLL |
54MHz (0 mW green) 38MHz (25 mW green) |
DDS3 | ch1 | homodyne phase |
SQ: 75 deg ASQ:140 deg |
Details:
ch1 |
ch1 |
CC light is not coupled into the OPO at all, after realignment
CC: TEM00=19.8mV; hom= 4.6mV+2.2mV; mode-matching: 74.4%.
SHG output power dropped to half, so we realign the incident 1064nm again, and now the green power before MZ is 126.7mW.
When the green light is well coupled into OPO, the GRMC and MZ are not easy to be locked. Seems to be interference from the feedback light.
When put the LO flip mirror back, the LO light is not coupled into the 2nd HOM PD at all.(but it worked yesterday)
We have to adjust the lens position to achieve HOM DC balance.
The BAB nonlinear gain today without green and with 25.1mW green again: (this time we have considered the background by blocking the BAB.)
without green, Max=240mV-6mV(BG)= 234mV ; OPO T:7.117
with 25.1mW green, Max=1390mV (Min=96mV) ;OPO T:7.119
SQ: 75 deg ASQ:140 deg |
Pump power (mW) | 0mW | 25.1mW |
OPO temperature (kOhm) | 7.117 | 7.119 |
P-Pol PLL frequency (MHz) | 270 | 190 |
BAB maximum (mV) | 234 | 1390 |
nonlinear gain | 1 | 5.94 |
SQ: 75 deg ASQ:140 deg |
cc1 cc2 can be locked smoothly.
cc1 error signal EPS1OUT Vpp=266mV.
cc2 error signal EPS1OUT drifts up and down over time in the range of 348mV, the Vpp is around 170mV.
The parameters we use for SR785 are:
128Hz~102.4kHz AC
FFT1 LogMag Hanning
Compute Avgs: Yes; type: Exp./cont; # Avgs: 100 ; Display Avg: RMS
SAVE: Display to Disk
SQ: 75 deg ASQ:140 deg |
SQ: 75 deg ASQ:140 degS |
[Chien-Ming, Hsun-Chung]
To check the alignment of green pump to OPO, we measure the nonlinear gain of BAB maximum in 1st HOM PD.
without green, Max=248mV (forgot to subtract the background); p-pol PLL 270MHz; OPO T:7.127
with 25.1mW green, Max=1540mV (Min=120mV) (also forgot to subtract the background) ; p-pol PLL 190MHz; OPO T:7.129
after two hours..
with 25.1mW green, Max=1200mV (Min=120mV) ;OPO T:7.129
I realign the green to OPO and adjust temp. of OPO
with 25.1mW green, Max=1580mV (Min=120mV) ;OPO T:7.123
without green, Max=248mV; OPO T:7.122
green power (mW) | 0mW | 25.1mW |
OPO temperature (kOhm) | 7.122 | 7.123 |
p pol PLL frequency (MHz) | 270 | 190 |
BAB maximum (mV) | 248* | 1580* |
nonlinear gain | 1 | 6.37 |
*This value does not check and subtract the background caused by p-pol.
To optimize HOM visibility, we realign the OPO output BAB mode matching at AMC.
Before: TEM00=164mV; hom=34mV+8mV+8mV ; 76.6% mode-matching
After: TEM00=218mV; hom=8mV+8mV ; 93.2% mode-matching
Realign the IRMC mode matching by monitoring the T of IRMC servo
Before: TEM00=920mV; hom=48mV+48mV; 90.6% mode-matching
After: TEM00=960mV; hom=56mV; 94.5% mode-matching
IRMC PDH signal and SHG PDH signal are using the same DDS channel to demodulate, so readjust the DDS phase from 110 to 140 degrees to reach a compromise. (Used 160 degrees to reach the best SHG PDH signal yesterday)
LO mode matching at AMC
TEM00=6.64V; hom=16mV ; 99.76% mode-matching
GRMC failed to lock again, checking the PDH signal and reloading the DDS2 to change the phase of GRMC demod. from 165 degrees to 200 degrees (Used 100 degrees yesterday).
7/25
[Chien-Ming, Hsun-Chung, Marc, Michael]
We realign the P-pol and BAB by using the steering mirrors before PBS
p-pol :
Before: TEM00=1.78V; hom=460mV+100mV ; 76% mode-matching
After: TEM00=1.92V; hom=160mV+160mV ; 85.7% mode-matching
BAB :
Before: TEM00=1.05V; hom=264mV+200mV+180mV+140mV+50mV=834mV ; 55.7% mode-matching
After: TEM00=1.17V; hom=232mV+200mV+200mV+140mV+50mV=822mV ; 58.7% mode-matching
To check the green pump to OPO alignment, we use BAB to optimize the nonlinear gain.
Without green, we phase lock the p-pol at 270MHz and tune the OPO temp. from 7.147 to 7.123.
we obtain the maximum BAB output by putting the power meter before homodyne BS.
The value is 1.78V - 900mV(background) = 880mV (however, power meter is set at 532nm)
When we want to introduce the green pump, we failed to lock GRMC and MZ, also the SHG output is not good.
So I realign the SHG input 1064nm beam,
Before:TEM00=1.58V; hom=1.28V; 55.2% mode-matching
After:TEM00= 354mV; hom:10mV; 97.3% mode-matching
Then I check the PDH signal of SHG and optimize it by changing the phase from 110 degrees to 160 degrees.
I successfully locked SHG. The power in front of MZ is 132.6 mW.
Then I check the PDH signal of GRMC, it's quite small, the S/N=280mV/50mV ~ 280mV/100mV.
Optimizing this signal by adjusting the mirror at the GRMC input reflected beam and then reloading the DDS2 to change the phase of GRMC demod. from 165 degrees to 100 degrees. We obtain the S/N= 126mV/15mV.
Finally, we successfully locked GRMC after adjusting the parameters of the servo.
After locking the MZ, the pump power corresponding to the MZ offset is as follows:
MZ offset | Green pump power |
4.0 | 15.7 mW |
4.1 | 21.6 mW |
4.2 | 27.0 mW |
4.3 | 32.3 mW |
4.4 | 37.8 mW |
4.5 | 43.4 mW |
4.6 | 49.0 mW |
4.7 | 54.6 mW |
4.8 | 59.6 mW |
4.9 | 64.9 mW |
5.0 | 70.1 mW |
I scanned LC1 and LC2 voltages from 0 to 25V with 0.02V step at 30degC.
Today I rotated LC1 by +90 deg to investigate the phase difference with respect to simulation.
I started measurement between 0 to 25V with 0.01V step at 30degC.
I tried to generate again polarization states but this time with 0.01V increment instead of the previously used 0.05V.
Because of TAMA visit I had to turn off the laser before the several days long measurement finished.
In any case, the results shows again good agreement with previous results and simulation and also higher polarization states coverage.
This result can be compared with simulation were I assumed for now same retardance vs voltage and extinction ratio equals to 1.
The rotation and retardation are shown in figures 1 and 2.
The retardation agrees really well with measurement if we wrap the measured retardation to positive values (fig 3).
Forthe rotation, it seems we have some kind of 90deg offset. I'm wondering if it could be because the first LC is rotated by -45deg instead of +45deg.
[Chien-Ming, Hsun-Chung, Marc, Michael]
First we found out that the power just after SHG is a bit lower than usual (measured 216mW) so we might have to tune its temperature.
Then, we found that MZ and GRMC were quite misaligned (especially in pitch).
We placed a power meter in transmission of GRMC to realign them.
First, the power had some periodic increase so we reduced the MZ scan period from 10 to 0.
Also, realigning it helped. After realignment we got :
1st arm TEM00 2.54V ; hom = 0.18V : 93% mode-matching
2nd arm TEM00 308mV ; 3 hom each = 6.4mV ; 94%
We tried to lock GRMC and MZ but the lock was not stable. So we had to increase the gains as follow:
increased GRMC gain to 3.1 from 1 ; increased MZ gain from 1 to 3
Finally, they locked and with MZ offset 4.54 we measured 44.7mW in transmission of GRMC in good agreement with the expected value.
After spending a lot of time generating polarization states and rotating the output polarizers I found out that this method is not too precise.
While the overall shape is similar to simulation, there are several points were the fitting is failing and gives polarization rotation 0 or 90 deg and retardation 45deg.
This is especially annoying as the measurement is far longer than expected (poor vi implementation limited measurement speed to 6Hz and need to rotate at least 36 times the ouput polarizer)...
In the end I removed the ouput polarizer and used the polarization camera readout.
THe measurement are agreeing now really well with simulations.
Note that in the simulation an arccos of the retardation limits the retardation to positive values only.
This result can be compared with simulation were I assumed for now same retardance vs voltage and extinction ratio equals to 1.
The rotation and retardation are shown in figures 1 and 2.
The retardation agrees really well with measurement if we wrap the measured retardation to positive values (fig 3).
Forthe rotation, it seems we have some kind of 90deg offset. I'm wondering if it could be because the first LC is rotated by -45deg instead of +45deg.
Marc, Pierre (remote)
Since the replacement of the EOM, we had difficulties to achieve automatic lock of IRMC.
This is mainly due to the fact that we're using 78MHz sidebands instead of 88MHz and had to change the PD in reflection of the IRMC for a non-resonant one.
First, the reflection signal of a scan of IRMC is different than before.
The base is at 1.38V and the TEM00 dip at 692mW while in the past the dip was negative.
This meant that we had to modify the pin P4 strap from pin 2-3 to pin 1-2.
Also, the threshold value was tweaked looking at this signal.
Then, the error signal was about 83mVpk while it was about 10 times more in the past.
Despite putting the servo gain to max I could not achieve lock (even though it was detected from the blinking LED).
I used the CCFC amplifier (32dB) and got an error signal about 2.2Vpk.
As the error signal slope and triangle scan slope are identical, the lock can be achieved with the switch INV.
I put a gain of 0.5 and could lock reliably the IRMC.
Marc, Michael
I wanted to check some of the multiplexer output diagnostics for the PLLs but the BEAT signal (the electrical signal from the fiber PD) had disappeared. We roughly checked the coupling into the CC fiber PD using the power meter and it seems reasonable (ML 4.34 mW, CC 0.64 mW, 1.05 mW out). The fiber PD electrical output have a large DC offset. I noticed that the PPol signal was oscillating at 45 kHz which I thought was a bit strange.
I eventually plugged the CC BEAT signal into the spectrum analyser with a wide setting, and noticed that the large peak corresponding to what was at the level of the previous measured CC BEAT (-11 dBm) had drifted quite far away from its usual frequency setting and the PPol BEAT peak was sitting at the previously mentioned 45 kHz. Both aux laser temperatures were quite far from the values listed in the wiki, so I readjusted them to the proper values and saw the good BEAT signals (CC -11 dBm, PPol -7 dBm).
However, the CC PLL lock is still unstable. Using MUXOUT on the CC PLL board, I could see that the digital lock detection was not particularly great. Digital lock detection outputs 1 when the phase delay between 3 cycles is less than 15 ns, and 0 when not. I noticed the output was not continuously 1 (i.e. has a lot of phase noise). It remained locked for about 4-5 minutes. I still need to figure out what the other diagnostic functions do.
The step voltage sub vi was fixed (misplaced connection) so I will spend the week end trying to generate and analyze 40 000 polarization states.
I'm scanning each LC voltage from 0 to 10V with 0.05V increment, at 30degC.
The output polarizer angle conversion is so that 298.25deg is our 0deg (ie cross polarizers configuration). The output polarizer will be rotated by 5deg increment.
Note that for tprevious measurement there is a -5deg offset with cross polarizer angle as it was assumed to be 293.25 deg.
I checked the CC PLL fast and slow control as described in 1205, with the only difference being that I used 21 MHz as the base signal. The purpose is to see how the PLL control loop responds to small modulations in a beat signal that is somewhat offset from the main LO.
For 1 MHz deviation, I saw DC offset ~ 50 mV (vs 10V for Yuhang) and no apparent AC signal. (fig 1)
For 1 kHz deviation, I saw DC offset ~ 50 mV and a somewhat messy signal with several frequency components. One is about 1.7 kHz (fig 2) and 100 Hz is highlighted in figure 3.
Seems like the fast loop is behaving a bit strangely
I also stared at the fast and slow loop oscilloscope signal for a few minutes each but didn't see much glitch behaviour. Probably not a conclusive test though.
[Marc, Shalika]
Labview modifications :
A new sub-vi has been created 'step voltage' that scan LC1 voltage from min to max voltage with user defined step. Once it reaches the maximum, the LC2 voltage is increased by the same step.
The time of each iteration was defined to have 1 data / step.
There is some little issue sometimes where Vlc2 > Vmax... Maybe du to our exit condition that does not take into account the finite numerical resolution we might have...
Controller issue
We found out that when the Kinessis software is started after Labview, the voltage controller of LC1 can not be used..
This is indicated in our vi front panel if the 'controller' LED is off.
This was the case for previous measurement and is now corrected.
First polarization state generation!!
Figures 1 and 2 are respectively measurement and simulation.
They agree quite well with each others!
We need to improve the fit by taking more measurement (was done every 10 deg rotation) and improve the simulation by taking the temperature and extinction ratio of LC1.
The step voltage sub vi was fixed (misplaced connection) so I will spend the week end trying to generate and analyze 40 000 polarization states.
I'm scanning each LC voltage from 0 to 10V with 0.05V increment, at 30degC.
The output polarizer angle conversion is so that 298.25deg is our 0deg (ie cross polarizers configuration). The output polarizer will be rotated by 5deg increment.
Note that for tprevious measurement there is a -5deg offset with cross polarizer angle as it was assumed to be 293.25 deg.
The goal is to scan the voltages of LC1 and LC2 and monitor the generated polarization state for each voltages couple.
LC1 fast axis is at 45deg of input polarizer and LC2 fast axis is at 0 deg of input polarizer.
The polarization state is scanned by rotating the output polarizer rotation by 5 deg increment from 0 deg (cross polarizer configuration) to 180 deg.
Labview modifications :
In order to somewhat efficiently scan our voltage space parameters, we modified the 'Sawtooth Formula Voltage Theta' subvi to scan our parameter space.
We can apply 1 sawtooth to each LC voltage controller with a user defined phase shift and frequency ratio. Below we used a 11:12 frequency ratio with 0.5 Hz frequency applied to LC1 (11/12*0.5Hz for LC2).
They are each doing 10 periods and then the output polarizer is rotated by 5 deg.
Analysis :
The data are stored in 2 files : one from rotation angle of 0 to 90 deg and one from 95 to 180 deg.
All the data are combined, voltages are rounded to 0.1V and for every couples (V_LC1,V_LC2) if there are at least 25 rotation angles values, the data are further processed. More precisely, the normalized transmitted power is fitted by P(theta) = a/2 * ( 1 + cos (2*ellipticity) + cos (2 * (theta - azimuth))
In the fit, the initial guess of the azimuth is given by the rotation angle with maximum transmitted power.
You can see some results in the attached figure (in total we have 1161 couples that meet our requirements).
next steps
I would like to replace the sawtooth scan by a step function scan to avoid the 0.1V rounding.
It could be useful to optimize our Lissajous scanning coefficients
In order to characterize the polarization states generated by our LCs, we can rotate the output polarizer while applying different voltages to our LC.
As our motorized rotator is from Thorlabs, we had to switch the position of our input and output polarizers.
We estimated the extinction ratio of our polarizers to be about 25 000 each.
Then, we calibrated again each LCs as shown in fig 1 and 2 for LC1 and 3 and 4 for LC2.
Their respective fast axis direction is 45.74 deg and 69.00 deg.
From figures 2 and 4, the maximum retardation at these angles is below 6 nm offset from 532 nm which indicates a misalignment of about 0.01 deg from the fast axis from simulation.
[Marc, Hirata, Sato, Takahashi]
This morning we opened PR chamber to take measurement of the in-vacuum Faraday Isolator. Indeed, the base plate has to be modified to be compatible with the new rotator.
Sato-san took precise measurements but it is something like 80mm along the beam direction and 90 mm in the transverse direction. In addition, we might have to move output polarizer holder. We have about 5 mm margin left and right (screw is at the center of the range) and the holder base is 25 mm from the edge of the FI baseplate towards the squeezer bench.
Also, because the BS oplevs are again giving strange signals, I realigned the green beam into the filter cavity before closing the chamber.
Moving PR I centered the beam on BS gate valve and got the beam on the first target (usual top-left of the hole). Then, I move both PR and BS to maintain the beam on the first target while recovering mostly the PR target.
Finally, I moved BS to get the beam on the 2nd target. At this position, I could confirm that both PR and BS coils actuators are working as expected.
The vacuum is now on-going.
[Marc Shalika]
For all these measurements we set up HWP and QWP at the input beam to have as linear as possible light from our polarization camera.
Then, we placed the input polarizer aligned with this linear polarizer as in fig1. We got aximuth angle ~ - 0.05+/-0.05 deg and ellipticity ~ -0.23+/-0.02 deg.
After installing the output cross polarizer we got azimuth angle ~ -0.025+/-0.05 deg and ellipticity ~ 0.06 +/-0.02 deg as in fig 2.
- Extinction ratio
Before installing the output polarizer we installed each LC successively. We rotated the LC to find the minimum and maximum transmission. Then we swept the LC voltage from 0 to 25V and computed the extinction ratio from the transmitted power normalized by the input power.
For LC2, we found the max and min positions were matching well the principal axes but it was not the case for LC1...
- Fast axis direction
As pointed out by Shalika, our previous estimations of the LC fast axis where really dependent on the fitting parameters range. It could come from the fact that using several sine harmonics in our fit biased our estimation. We decided to use a different formula : P_trans = a * sin(2(theta-theta0))^2 with theta the rotation angle of the LC.
We swept the LC voltages at various rotation angles covering more than 90 deg as in fig4 for LC2. From the fit we could extract the fast axis direction of 12.18 deg. This is in really good agreement with our 'by hand' estimation of 11.43deg.
All the swept results are reported in fig5. It can be seen that we get the usual retardance varying from 17nm to 989 nm as a function of applied voltage for every rotation angle except when the LC fast axis is close to the input polarization direction. In that case the maximum retardance is only ~650 nm while smallest one is increased to ~ 60 nm.
We simulated a rotating LC inside cross polarizers. Input polarization and polarizers are assumed perfect but we added by hand a backgroud power of 22.12 nW as measured in cross polarizers without LC. The LC voltage response is coming from the fit of the value at 45deg rotation wrt input polarization direction and takes into account measured extinction ratio. Results are reported in fig6 and agree really well with our measurement (especially at low voltage). An offset of 0.75 deg creates a maximum retardance of 657nm!
For LC1, we repeated the same measurement and measured fast axis direction of 145.05 deg.
- Temperature effect
All these measurements are performed at 30 degC nominal value and we typically see variation of less than 0.1 degC.
We measured the retardance at 0V while changing the LC temperature and results are attached in fig 7.
For LC2 we measured a change of 9.81 nm / degC while for LC1 7.84 nm/degC.
I looked at the CC PLL again. It still seems to be quite unstable. There was one occasion where I could get it to lock for more than 10 minutes. During phase noise measurement I saw that the noise was glitch type rather than stationary, where the noise floor could be at the level of previous measurements but quickly gets pushed up by an impulse excitation.
This time I measured the phase noise by providing 7 MHz as the local oscillator from DDS3 DAC0 (PPol LO @ 9 dBm), which corresponds to the usual operation frequency of the CC PLL beat note (7MHz at "-23.5" dBm on the "17dB more is measured " spectrum analyser, DDS signal to CC LO is 21 MHz). Changing the LO offset to both 6.9990 MHz and 7.0001 MHz (+/-100 Hz offset) resulted in a corresponding 100 Hz signal output from the mixer. On the oscilloscope I can still quite frequently see the signal glitches.
I tried twice to get the phase noise spectrum but it seems the signal is not cooperating. The noise floor gets raised quite high by random noise impulses (figure 1). Qualitatively, the noise floor should be about at the level of the other measurements.
Figure 2 shows the Apk calibration to go from Vrms to rad/rtHz. Last time it was 0.0086 mV, now it is 0.0078.
Figure 3 and 4 show some feedback on the CC spectrum analyser at 14 and 28 MHz. The CC PLL is going to a T attached on the spectrum analyser. The other exit of the T then goes to the RF port of a mixer. When disconnect the mixer the 14/28... peaks go away.
Figure 5 shows the period of 2 wavelengths when the local oscillator frequency is offset to 7.000 100 MHz in DDS3 DAC3. So a 100 Hz difference between DDS and beatnote produces the 100 Hz signal as expected.
The humidity in the west side of the tunnel is increasing (85%). I set the fan in the tunnel (photo). It is working during this rain season.
A parameter for a phase shifted voltage waveform between the two LCs has also been added.
The VI is completely ready for use. The speed of VI (with all data saving) is 40Hz.